unsupervised image classification: Topics by Science.gov

Sample records for unsupervised image classification

On the Implementation of a Land Cover Classification System for SAR Images Using Khoros

NASA Technical Reports Server (NTRS)

Medina Revera, Edwin J.; Espinosa, Ramon Vasquez

1997-01-01

The Synthetic Aperture Radar (SAR) sensor is widely used to record data about the ground under all atmospheric conditions. The SAR acquired images have very good resolution which necessitates the development of a classification system that process the SAR images to extract useful information for different applications. In this work, a complete system for the land cover classification was designed and programmed using the Khoros, a data flow visual language environment, taking full advantages of the polymorphic data services that it provides. Image analysis was applied to SAR images to improve and automate the processes of recognition and classification of the different regions like mountains and lakes. Both unsupervised and supervised classification utilities were used. The unsupervised classification routines included the use of several Classification/Clustering algorithms like the K-means, ISO2, Weighted Minimum Distance, and the Localized Receptive Field (LRF) training/classifier. Different texture analysis approaches such as Invariant Moments, Fractal Dimension and Second Order statistics were implemented for supervised classification of the images. The results and conclusions for SAR image classification using the various unsupervised and supervised procedures are presented based on their accuracy and performance.
Classification of earth terrain using polarimetric synthetic aperture radar images

NASA Technical Reports Server (NTRS)

Lim, H. H.; Swartz, A. A.; Yueh, H. A.; Kong, J. A.; Shin, R. T.; Van Zyl, J. J.

1989-01-01

Supervised and unsupervised classification techniques are developed and used to classify the earth terrain components from SAR polarimetric images of San Francisco Bay and Traverse City, Michigan. The supervised techniques include the Bayes classifiers, normalized polarimetric classification, and simple feature classification using discriminates such as the absolute and normalized magnitude response of individual receiver channel returns and the phase difference between receiver channels. An algorithm is developed as an unsupervised technique which classifies terrain elements based on the relationship between the orientation angle and the handedness of the transmitting and receiving polariation states. It is found that supervised classification produces the best results when accurate classifier training data are used, while unsupervised classification may be applied when training data are not available.
Training strategy for convolutional neural networks in pedestrian gender classification

NASA Astrophysics Data System (ADS)

Ng, Choon-Boon; Tay, Yong-Haur; Goi, Bok-Min

2017-06-01

In this work, we studied a strategy for training a convolutional neural network in pedestrian gender classification with limited amount of labeled training data. Unsupervised learning by k-means clustering on pedestrian images was used to learn the filters to initialize the first layer of the network. As a form of pre-training, supervised learning for the related task of pedestrian classification was performed. Finally, the network was fine-tuned for gender classification. We found that this strategy improved the network's generalization ability in gender classification, achieving better test results when compared to random weights initialization and slightly more beneficial than merely initializing the first layer filters by unsupervised learning. This shows that unsupervised learning followed by pre-training with pedestrian images is an effective strategy to learn useful features for pedestrian gender classification.
Comparing supervised and unsupervised multiresolution segmentation approaches for extracting buildings from very high resolution imagery.

PubMed

Belgiu, Mariana; Dr Guţ, Lucian

2014-10-01

Although multiresolution segmentation (MRS) is a powerful technique for dealing with very high resolution imagery, some of the image objects that it generates do not match the geometries of the target objects, which reduces the classification accuracy. MRS can, however, be guided to produce results that approach the desired object geometry using either supervised or unsupervised approaches. Although some studies have suggested that a supervised approach is preferable, there has been no comparative evaluation of these two approaches. Therefore, in this study, we have compared supervised and unsupervised approaches to MRS. One supervised and two unsupervised segmentation methods were tested on three areas using QuickBird and WorldView-2 satellite imagery. The results were assessed using both segmentation evaluation methods and an accuracy assessment of the resulting building classifications. Thus, differences in the geometries of the image objects and in the potential to achieve satisfactory thematic accuracies were evaluated. The two approaches yielded remarkably similar classification results, with overall accuracies ranging from 82% to 86%. The performance of one of the unsupervised methods was unexpectedly similar to that of the supervised method; they identified almost identical scale parameters as being optimal for segmenting buildings, resulting in very similar geometries for the resulting image objects. The second unsupervised method produced very different image objects from the supervised method, but their classification accuracies were still very similar. The latter result was unexpected because, contrary to previously published findings, it suggests a high degree of independence between the segmentation results and classification accuracy. The results of this study have two important implications. The first is that object-based image analysis can be automated without sacrificing classification accuracy, and the second is that the previously accepted idea that classification is dependent on segmentation is challenged by our unexpected results, casting doubt on the value of pursuing 'optimal segmentation'. Our results rather suggest that as long as under-segmentation remains at acceptable levels, imperfections in segmentation can be ruled out, so that a high level of classification accuracy can still be achieved.
Semi-supervised classification tool for DubaiSat-2 multispectral imagery

NASA Astrophysics Data System (ADS)

Al-Mansoori, Saeed

2015-10-01

This paper addresses a semi-supervised classification tool based on a pixel-based approach of the multi-spectral satellite imagery. There are not many studies demonstrating such algorithm for the multispectral images, especially when the image consists of 4 bands (Red, Green, Blue and Near Infrared) as in DubaiSat-2 satellite images. The proposed approach utilizes both unsupervised and supervised classification schemes sequentially to identify four classes in the image, namely, water bodies, vegetation, land (developed and undeveloped areas) and paved areas (i.e. roads). The unsupervised classification concept is applied to identify two classes; water bodies and vegetation, based on a well-known index that uses the distinct wavelengths of visible and near-infrared sunlight that is absorbed and reflected by the plants to identify the classes; this index parameter is called "Normalized Difference Vegetation Index (NDVI)". Afterward, the supervised classification is performed by selecting training homogenous samples for roads and land areas. Here, a precise selection of training samples plays a vital role in the classification accuracy. Post classification is finally performed to enhance the classification accuracy, where the classified image is sieved, clumped and filtered before producing final output. Overall, the supervised classification approach produced higher accuracy than the unsupervised method. This paper shows some current preliminary research results which point out the effectiveness of the proposed technique in a virtual perspective.
Unsupervised feature learning for autonomous rock image classification

NASA Astrophysics Data System (ADS)

Shu, Lei; McIsaac, Kenneth; Osinski, Gordon R.; Francis, Raymond

2017-09-01

Autonomous rock image classification can enhance the capability of robots for geological detection and enlarge the scientific returns, both in investigation on Earth and planetary surface exploration on Mars. Since rock textural images are usually inhomogeneous and manually hand-crafting features is not always reliable, we propose an unsupervised feature learning method to autonomously learn the feature representation for rock images. In our tests, rock image classification using the learned features shows that the learned features can outperform manually selected features. Self-taught learning is also proposed to learn the feature representation from a large database of unlabelled rock images of mixed class. The learned features can then be used repeatedly for classification of any subclass. This takes advantage of the large dataset of unlabelled rock images and learns a general feature representation for many kinds of rocks. We show experimental results supporting the feasibility of self-taught learning on rock images.
Unsupervised hierarchical partitioning of hyperspectral images: application to marine algae identification

NASA Astrophysics Data System (ADS)

Chen, B.; Chehdi, K.; De Oliveria, E.; Cariou, C.; Charbonnier, B.

2015-10-01

In this paper a new unsupervised top-down hierarchical classification method to partition airborne hyperspectral images is proposed. The unsupervised approach is preferred because the difficulty of area access and the human and financial resources required to obtain ground truth data, constitute serious handicaps especially over large areas which can be covered by airborne or satellite images. The developed classification approach allows i) a successive partitioning of data into several levels or partitions in which the main classes are first identified, ii) an estimation of the number of classes automatically at each level without any end user help, iii) a nonsystematic subdivision of all classes of a partition Pj to form a partition Pj+1, iv) a stable partitioning result of the same data set from one run of the method to another. The proposed approach was validated on synthetic and real hyperspectral images related to the identification of several marine algae species. In addition to highly accurate and consistent results (correct classification rate over 99%), this approach is completely unsupervised. It estimates at each level, the optimal number of classes and the final partition without any end user intervention.
Novel Histogram Based Unsupervised Classification Technique to Determine Natural Classes From Biophysically Relevant Fit Parameters to Hyperspectral Data

DOE PAGES

McCann, Cooper; Repasky, Kevin S.; Morin, Mikindra; ...

2017-05-23

Hyperspectral image analysis has benefited from an array of methods that take advantage of the increased spectral depth compared to multispectral sensors; however, the focus of these developments has been on supervised classification methods. Lack of a priori knowledge regarding land cover characteristics can make unsupervised classification methods preferable under certain circumstances. An unsupervised classification technique is presented in this paper that utilizes physically relevant basis functions to model the reflectance spectra. These fit parameters used to generate the basis functions allow clustering based on spectral characteristics rather than spectral channels and provide both noise and data reduction. Histogram splittingmore » of the fit parameters is then used as a means of producing an unsupervised classification. Unlike current unsupervised classification techniques that rely primarily on Euclidian distance measures to determine similarity, the unsupervised classification technique uses the natural splitting of the fit parameters associated with the basis functions creating clusters that are similar in terms of physical parameters. The data set used in this work utilizes the publicly available data collected at Indian Pines, Indiana. This data set provides reference data allowing for comparisons of the efficacy of different unsupervised data analysis. The unsupervised histogram splitting technique presented in this paper is shown to be better than the standard unsupervised ISODATA clustering technique with an overall accuracy of 34.3/19.0% before merging and 40.9/39.2% after merging. Finally, this improvement is also seen as an improvement of kappa before/after merging of 24.8/30.5 for the histogram splitting technique compared to 15.8/28.5 for ISODATA.« less
Novel Histogram Based Unsupervised Classification Technique to Determine Natural Classes From Biophysically Relevant Fit Parameters to Hyperspectral Data

DOE Office of Scientific and Technical Information (OSTI.GOV)

McCann, Cooper; Repasky, Kevin S.; Morin, Mikindra

Hyperspectral image analysis has benefited from an array of methods that take advantage of the increased spectral depth compared to multispectral sensors; however, the focus of these developments has been on supervised classification methods. Lack of a priori knowledge regarding land cover characteristics can make unsupervised classification methods preferable under certain circumstances. An unsupervised classification technique is presented in this paper that utilizes physically relevant basis functions to model the reflectance spectra. These fit parameters used to generate the basis functions allow clustering based on spectral characteristics rather than spectral channels and provide both noise and data reduction. Histogram splittingmore » of the fit parameters is then used as a means of producing an unsupervised classification. Unlike current unsupervised classification techniques that rely primarily on Euclidian distance measures to determine similarity, the unsupervised classification technique uses the natural splitting of the fit parameters associated with the basis functions creating clusters that are similar in terms of physical parameters. The data set used in this work utilizes the publicly available data collected at Indian Pines, Indiana. This data set provides reference data allowing for comparisons of the efficacy of different unsupervised data analysis. The unsupervised histogram splitting technique presented in this paper is shown to be better than the standard unsupervised ISODATA clustering technique with an overall accuracy of 34.3/19.0% before merging and 40.9/39.2% after merging. Finally, this improvement is also seen as an improvement of kappa before/after merging of 24.8/30.5 for the histogram splitting technique compared to 15.8/28.5 for ISODATA.« less
Unsupervised Cryo-EM Data Clustering through Adaptively Constrained K-Means Algorithm

PubMed Central

Xu, Yaofang; Wu, Jiayi; Yin, Chang-Cheng; Mao, Youdong

2016-01-01

In single-particle cryo-electron microscopy (cryo-EM), K-means clustering algorithm is widely used in unsupervised 2D classification of projection images of biological macromolecules. 3D ab initio reconstruction requires accurate unsupervised classification in order to separate molecular projections of distinct orientations. Due to background noise in single-particle images and uncertainty of molecular orientations, traditional K-means clustering algorithm may classify images into wrong classes and produce classes with a large variation in membership. Overcoming these limitations requires further development on clustering algorithms for cryo-EM data analysis. We propose a novel unsupervised data clustering method building upon the traditional K-means algorithm. By introducing an adaptive constraint term in the objective function, our algorithm not only avoids a large variation in class sizes but also produces more accurate data clustering. Applications of this approach to both simulated and experimental cryo-EM data demonstrate that our algorithm is a significantly improved alterative to the traditional K-means algorithm in single-particle cryo-EM analysis. PMID:27959895
Unsupervised Cryo-EM Data Clustering through Adaptively Constrained K-Means Algorithm.

PubMed

Xu, Yaofang; Wu, Jiayi; Yin, Chang-Cheng; Mao, Youdong

2016-01-01

In single-particle cryo-electron microscopy (cryo-EM), K-means clustering algorithm is widely used in unsupervised 2D classification of projection images of biological macromolecules. 3D ab initio reconstruction requires accurate unsupervised classification in order to separate molecular projections of distinct orientations. Due to background noise in single-particle images and uncertainty of molecular orientations, traditional K-means clustering algorithm may classify images into wrong classes and produce classes with a large variation in membership. Overcoming these limitations requires further development on clustering algorithms for cryo-EM data analysis. We propose a novel unsupervised data clustering method building upon the traditional K-means algorithm. By introducing an adaptive constraint term in the objective function, our algorithm not only avoids a large variation in class sizes but also produces more accurate data clustering. Applications of this approach to both simulated and experimental cryo-EM data demonstrate that our algorithm is a significantly improved alterative to the traditional K-means algorithm in single-particle cryo-EM analysis.
A comparative analysis of pixel- and object-based detection of landslides from very high-resolution images

NASA Astrophysics Data System (ADS)

Keyport, Ren N.; Oommen, Thomas; Martha, Tapas R.; Sajinkumar, K. S.; Gierke, John S.

2018-02-01

A comparative analysis of landslides detected by pixel-based and object-oriented analysis (OOA) methods was performed using very high-resolution (VHR) remotely sensed aerial images for the San Juan La Laguna, Guatemala, which witnessed widespread devastation during the 2005 Hurricane Stan. A 3-band orthophoto of 0.5 m spatial resolution together with a 115 field-based landslide inventory were used for the analysis. A binary reference was assigned with a zero value for landslide and unity for non-landslide pixels. The pixel-based analysis was performed using unsupervised classification, which resulted in 11 different trial classes. Detection of landslides using OOA includes 2-step K-means clustering to eliminate regions based on brightness; elimination of false positives using object properties such as rectangular fit, compactness, length/width ratio, mean difference of objects, and slope angle. Both overall accuracy and F-score for OOA methods outperformed pixel-based unsupervised classification methods in both landslide and non-landslide classes. The overall accuracy for OOA and pixel-based unsupervised classification was 96.5% and 94.3%, respectively, whereas the best F-score for landslide identification for OOA and pixel-based unsupervised methods: were 84.3% and 77.9%, respectively.Results indicate that the OOA is able to identify the majority of landslides with a few false positive when compared to pixel-based unsupervised classification.
Massively parallel unsupervised single-particle cryo-EM data clustering via statistical manifold learning

PubMed Central

Wu, Jiayi; Ma, Yong-Bei; Congdon, Charles; Brett, Bevin; Chen, Shuobing; Xu, Yaofang; Ouyang, Qi

2017-01-01

Structural heterogeneity in single-particle cryo-electron microscopy (cryo-EM) data represents a major challenge for high-resolution structure determination. Unsupervised classification may serve as the first step in the assessment of structural heterogeneity. However, traditional algorithms for unsupervised classification, such as K-means clustering and maximum likelihood optimization, may classify images into wrong classes with decreasing signal-to-noise-ratio (SNR) in the image data, yet demand increased computational costs. Overcoming these limitations requires further development of clustering algorithms for high-performance cryo-EM data processing. Here we introduce an unsupervised single-particle clustering algorithm derived from a statistical manifold learning framework called generative topographic mapping (GTM). We show that unsupervised GTM clustering improves classification accuracy by about 40% in the absence of input references for data with lower SNRs. Applications to several experimental datasets suggest that our algorithm can detect subtle structural differences among classes via a hierarchical clustering strategy. After code optimization over a high-performance computing (HPC) environment, our software implementation was able to generate thousands of reference-free class averages within hours in a massively parallel fashion, which allows a significant improvement on ab initio 3D reconstruction and assists in the computational purification of homogeneous datasets for high-resolution visualization. PMID:28786986
Massively parallel unsupervised single-particle cryo-EM data clustering via statistical manifold learning.

PubMed

Wu, Jiayi; Ma, Yong-Bei; Congdon, Charles; Brett, Bevin; Chen, Shuobing; Xu, Yaofang; Ouyang, Qi; Mao, Youdong

2017-01-01

Structural heterogeneity in single-particle cryo-electron microscopy (cryo-EM) data represents a major challenge for high-resolution structure determination. Unsupervised classification may serve as the first step in the assessment of structural heterogeneity. However, traditional algorithms for unsupervised classification, such as K-means clustering and maximum likelihood optimization, may classify images into wrong classes with decreasing signal-to-noise-ratio (SNR) in the image data, yet demand increased computational costs. Overcoming these limitations requires further development of clustering algorithms for high-performance cryo-EM data processing. Here we introduce an unsupervised single-particle clustering algorithm derived from a statistical manifold learning framework called generative topographic mapping (GTM). We show that unsupervised GTM clustering improves classification accuracy by about 40% in the absence of input references for data with lower SNRs. Applications to several experimental datasets suggest that our algorithm can detect subtle structural differences among classes via a hierarchical clustering strategy. After code optimization over a high-performance computing (HPC) environment, our software implementation was able to generate thousands of reference-free class averages within hours in a massively parallel fashion, which allows a significant improvement on ab initio 3D reconstruction and assists in the computational purification of homogeneous datasets for high-resolution visualization.
A comparison of unsupervised classification procedures on LANDSAT MSS data for an area of complex surface conditions in Basilicata, Southern Italy

NASA Technical Reports Server (NTRS)

Justice, C.; Townshend, J. (Principal Investigator)

1981-01-01

Two unsupervised classification procedures were applied to ratioed and unratioed LANDSAT multispectral scanner data of an area of spatially complex vegetation and terrain. An objective accuracy assessment was undertaken on each classification and comparison was made of the classification accuracies. The two unsupervised procedures use the same clustering algorithm. By on procedure the entire area is clustered and by the other a representative sample of the area is clustered and the resulting statistics are extrapolated to the remaining area using a maximum likelihood classifier. Explanation is given of the major steps in the classification procedures including image preprocessing; classification; interpretation of cluster classes; and accuracy assessment. Of the four classifications undertaken, the monocluster block approach on the unratioed data gave the highest accuracy of 80% for five coarse cover classes. This accuracy was increased to 84% by applying a 3 x 3 contextual filter to the classified image. A detailed description and partial explanation is provided for the major misclassification. The classification of the unratioed data produced higher percentage accuracies than for the ratioed data and the monocluster block approach gave higher accuracies than clustering the entire area. The moncluster block approach was additionally the most economical in terms of computing time.
An unsupervised technique for optimal feature selection in attribute profiles for spectral-spatial classification of hyperspectral images

NASA Astrophysics Data System (ADS)

Bhardwaj, Kaushal; Patra, Swarnajyoti

2018-04-01

Inclusion of spatial information along with spectral features play a significant role in classification of remote sensing images. Attribute profiles have already proved their ability to represent spatial information. In order to incorporate proper spatial information, multiple attributes are required and for each attribute large profiles need to be constructed by varying the filter parameter values within a wide range. Thus, the constructed profiles that represent spectral-spatial information of an hyperspectral image have huge dimension which leads to Hughes phenomenon and increases computational burden. To mitigate these problems, this work presents an unsupervised feature selection technique that selects a subset of filtered image from the constructed high dimensional multi-attribute profile which are sufficiently informative to discriminate well among classes. In this regard the proposed technique exploits genetic algorithms (GAs). The fitness function of GAs are defined in an unsupervised way with the help of mutual information. The effectiveness of the proposed technique is assessed using one-against-all support vector machine classifier. The experiments conducted on three hyperspectral data sets show the robustness of the proposed method in terms of computation time and classification accuracy.
Shadow detection and removal in RGB VHR images for land use unsupervised classification

NASA Astrophysics Data System (ADS)

Movia, A.; Beinat, A.; Crosilla, F.

2016-09-01

Nowadays, high resolution aerial images are widely available thanks to the diffusion of advanced technologies such as UAVs (Unmanned Aerial Vehicles) and new satellite missions. Although these developments offer new opportunities for accurate land use analysis and change detection, cloud and terrain shadows actually limit benefits and possibilities of modern sensors. Focusing on the problem of shadow detection and removal in VHR color images, the paper proposes new solutions and analyses how they can enhance common unsupervised classification procedures for identifying land use classes related to the CO2 absorption. To this aim, an improved fully automatic procedure has been developed for detecting image shadows using exclusively RGB color information, and avoiding user interaction. Results show a significant accuracy enhancement with respect to similar methods using RGB based indexes. Furthermore, novel solutions derived from Procrustes analysis have been applied to remove shadows and restore brightness in the images. In particular, two methods implementing the so called "anisotropic Procrustes" and the "not-centered oblique Procrustes" algorithms have been developed and compared with the linear correlation correction method based on the Cholesky decomposition. To assess how shadow removal can enhance unsupervised classifications, results obtained with classical methods such as k-means, maximum likelihood, and self-organizing maps, have been compared to each other and with a supervised clustering procedure.
Emotional textile image classification based on cross-domain convolutional sparse autoencoders with feature selection

NASA Astrophysics Data System (ADS)

Li, Zuhe; Fan, Yangyu; Liu, Weihua; Yu, Zeqi; Wang, Fengqin

2017-01-01

We aim to apply sparse autoencoder-based unsupervised feature learning to emotional semantic analysis for textile images. To tackle the problem of limited training data, we present a cross-domain feature learning scheme for emotional textile image classification using convolutional autoencoders. We further propose a correlation-analysis-based feature selection method for the weights learned by sparse autoencoders to reduce the number of features extracted from large size images. First, we randomly collect image patches on an unlabeled image dataset in the source domain and learn local features with a sparse autoencoder. We then conduct feature selection according to the correlation between different weight vectors corresponding to the autoencoder's hidden units. We finally adopt a convolutional neural network including a pooling layer to obtain global feature activations of textile images in the target domain and send these global feature vectors into logistic regression models for emotional image classification. The cross-domain unsupervised feature learning method achieves 65% to 78% average accuracy in the cross-validation experiments corresponding to eight emotional categories and performs better than conventional methods. Feature selection can reduce the computational cost of global feature extraction by about 50% while improving classification performance.
Unsupervised universal steganalyzer for high-dimensional steganalytic features

NASA Astrophysics Data System (ADS)

Hou, Xiaodan; Zhang, Tao

2016-11-01

The research in developing steganalytic features has been highly successful. These features are extremely powerful when applied to supervised binary classification problems. However, they are incompatible with unsupervised universal steganalysis because the unsupervised method cannot distinguish embedding distortion from varying levels of noises caused by cover variation. This study attempts to alleviate the problem by introducing similarity retrieval of image statistical properties (SRISP), with the specific aim of mitigating the effect of cover variation on the existing steganalytic features. First, cover images with some statistical properties similar to those of a given test image are searched from a retrieval cover database to establish an aided sample set. Then, unsupervised outlier detection is performed on a test set composed of the given test image and its aided sample set to determine the type (cover or stego) of the given test image. Our proposed framework, called SRISP-aided unsupervised outlier detection, requires no training. Thus, it does not suffer from model mismatch mess. Compared with prior unsupervised outlier detectors that do not consider SRISP, the proposed framework not only retains the universality but also exhibits superior performance when applied to high-dimensional steganalytic features.
MARTA GANs: Unsupervised Representation Learning for Remote Sensing Image Classification

NASA Astrophysics Data System (ADS)

Lin, Daoyu; Fu, Kun; Wang, Yang; Xu, Guangluan; Sun, Xian

2017-11-01

With the development of deep learning, supervised learning has frequently been adopted to classify remotely sensed images using convolutional networks (CNNs). However, due to the limited amount of labeled data available, supervised learning is often difficult to carry out. Therefore, we proposed an unsupervised model called multiple-layer feature-matching generative adversarial networks (MARTA GANs) to learn a representation using only unlabeled data. MARTA GANs consists of both a generative model $G$ and a discriminative model $D$. We treat $D$ as a feature extractor. To fit the complex properties of remote sensing data, we use a fusion layer to merge the mid-level and global features. $G$ can produce numerous images that are similar to the training data; therefore, $D$ can learn better representations of remotely sensed images using the training data provided by $G$. The classification results on two widely used remote sensing image databases show that the proposed method significantly improves the classification performance compared with other state-of-the-art methods.

Evaluation of solar angle variation over digital processing of LANDSAT imagery. [Brazil

NASA Technical Reports Server (NTRS)

Parada, N. D. J. (Principal Investigator); Novo, E. M. L. M.

1984-01-01

The effects of the seasonal variation of illumination over digital processing of LANDSAT images are evaluated. Original images are transformed by means of digital filtering to enhance their spatial features. The resulting images are used to obtain an unsupervised classification of relief units. After defining relief classes, which are supposed to be spectrally different, topographic variables (declivity, altitude, relief range and slope length) are used to identify the true relief units existing on the ground. The samples are also clustered by means of an unsupervised classification option. The results obtained for each LANDSAT overpass are compared. Digital processing is highly affected by illumination geometry. There is no correspondence between relief units as defined by spectral features and those resulting from topographic features.
An automatic taxonomy of galaxy morphology using unsupervised machine learning

NASA Astrophysics Data System (ADS)

Hocking, Alex; Geach, James E.; Sun, Yi; Davey, Neil

2018-01-01

We present an unsupervised machine learning technique that automatically segments and labels galaxies in astronomical imaging surveys using only pixel data. Distinct from previous unsupervised machine learning approaches used in astronomy we use no pre-selection or pre-filtering of target galaxy type to identify galaxies that are similar. We demonstrate the technique on the Hubble Space Telescope (HST) Frontier Fields. By training the algorithm using galaxies from one field (Abell 2744) and applying the result to another (MACS 0416.1-2403), we show how the algorithm can cleanly separate early and late type galaxies without any form of pre-directed training for what an 'early' or 'late' type galaxy is. We then apply the technique to the HST Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) fields, creating a catalogue of approximately 60 000 classifications. We show how the automatic classification groups galaxies of similar morphological (and photometric) type and make the classifications public via a catalogue, a visual catalogue and galaxy similarity search. We compare the CANDELS machine-based classifications to human-classifications from the Galaxy Zoo: CANDELS project. Although there is not a direct mapping between Galaxy Zoo and our hierarchical labelling, we demonstrate a good level of concordance between human and machine classifications. Finally, we show how the technique can be used to identify rarer objects and present lensed galaxy candidates from the CANDELS imaging.
A new tool for supervised classification of satellite images available on web servers: Google Maps as a case study

NASA Astrophysics Data System (ADS)

García-Flores, Agustín.; Paz-Gallardo, Abel; Plaza, Antonio; Li, Jun

2016-10-01

This paper describes a new web platform dedicated to the classification of satellite images called Hypergim. The current implementation of this platform enables users to perform classification of satellite images from any part of the world thanks to the worldwide maps provided by Google Maps. To perform this classification, Hypergim uses unsupervised algorithms like Isodata and K-means. Here, we present an extension of the original platform in which we adapt Hypergim in order to use supervised algorithms to improve the classification results. This involves a significant modification of the user interface, providing the user with a way to obtain samples of classes present in the images to use in the training phase of the classification process. Another main goal of this development is to improve the runtime of the image classification process. To achieve this goal, we use a parallel implementation of the Random Forest classification algorithm. This implementation is a modification of the well-known CURFIL software package. The use of this type of algorithms to perform image classification is widespread today thanks to its precision and ease of training. The actual implementation of Random Forest was developed using CUDA platform, which enables us to exploit the potential of several models of NVIDIA graphics processing units using them to execute general purpose computing tasks as image classification algorithms. As well as CUDA, we use other parallel libraries as Intel Boost, taking advantage of the multithreading capabilities of modern CPUs. To ensure the best possible results, the platform is deployed in a cluster of commodity graphics processing units (GPUs), so that multiple users can use the tool in a concurrent way. The experimental results indicate that this new algorithm widely outperform the previous unsupervised algorithms implemented in Hypergim, both in runtime as well as precision of the actual classification of the images.
Unsupervised learning of discriminative edge measures for vehicle matching between nonoverlapping cameras.

PubMed

Shan, Ying; Sawhney, Harpreet S; Kumar, Rakesh

2008-04-01

This paper proposes a novel unsupervised algorithm learning discriminative features in the context of matching road vehicles between two non-overlapping cameras. The matching problem is formulated as a same-different classification problem, which aims to compute the probability of vehicle images from two distinct cameras being from the same vehicle or different vehicle(s). We employ a novel measurement vector that consists of three independent edge-based measures and their associated robust measures computed from a pair of aligned vehicle edge maps. The weight of each measure is determined by an unsupervised learning algorithm that optimally separates the same-different classes in the combined measurement space. This is achieved with a weak classification algorithm that automatically collects representative samples from same-different classes, followed by a more discriminative classifier based on Fisher' s Linear Discriminants and Gibbs Sampling. The robustness of the match measures and the use of unsupervised discriminant analysis in the classification ensures that the proposed method performs consistently in the presence of missing/false features, temporally and spatially changing illumination conditions, and systematic misalignment caused by different camera configurations. Extensive experiments based on real data of over 200 vehicles at different times of day demonstrate promising results.
Analysis On Land Cover In Municipality Of Malang With Landsat 8 Image Through Unsupervised Classification

NASA Astrophysics Data System (ADS)

Nahari, R. V.; Alfita, R.

2018-01-01

Remote sensing technology has been widely used in the geographic information system in order to obtain data more quickly, accurately and affordably. One of the advantages of using remote sensing imagery (satellite imagery) is to analyze land cover and land use. Satellite image data used in this study were images from the Landsat 8 satellite combined with the data from the Municipality of Malang government. The satellite image was taken in July 2016. Furthermore, the method used in this study was unsupervised classification. Based on the analysis towards the satellite images and field observations, 29% of the land in the Municipality of Malang was plantation, 22% of the area was rice field, 12% was residential area, 10% was land with shrubs, and the remaining 2% was water (lake/reservoir). The shortcoming of the methods was 25% of the land in the area was unidentified because it was covered by cloud. It is expected that future researchers involve cloud removal processing to minimize unidentified area.
Hyperspectral image segmentation using a cooperative nonparametric approach

NASA Astrophysics Data System (ADS)

Taher, Akar; Chehdi, Kacem; Cariou, Claude

2013-10-01

In this paper a new unsupervised nonparametric cooperative and adaptive hyperspectral image segmentation approach is presented. The hyperspectral images are partitioned band by band in parallel and intermediate classification results are evaluated and fused, to get the final segmentation result. Two unsupervised nonparametric segmentation methods are used in parallel cooperation, namely the Fuzzy C-means (FCM) method, and the Linde-Buzo-Gray (LBG) algorithm, to segment each band of the image. The originality of the approach relies firstly on its local adaptation to the type of regions in an image (textured, non-textured), and secondly on the introduction of several levels of evaluation and validation of intermediate segmentation results before obtaining the final partitioning of the image. For the management of similar or conflicting results issued from the two classification methods, we gradually introduced various assessment steps that exploit the information of each spectral band and its adjacent bands, and finally the information of all the spectral bands. In our approach, the detected textured and non-textured regions are treated separately from feature extraction step, up to the final classification results. This approach was first evaluated on a large number of monocomponent images constructed from the Brodatz album. Then it was evaluated on two real applications using a respectively multispectral image for Cedar trees detection in the region of Baabdat (Lebanon) and a hyperspectral image for identification of invasive and non invasive vegetation in the region of Cieza (Spain). A correct classification rate (CCR) for the first application is over 97% and for the second application the average correct classification rate (ACCR) is over 99%.
Cluster Method Analysis of K. S. C. Image

NASA Technical Reports Server (NTRS)

Rodriguez, Joe, Jr.; Desai, M.

1997-01-01

Information obtained from satellite-based systems has moved to the forefront as a method in the identification of many land cover types. Identification of different land features through remote sensing is an effective tool for regional and global assessment of geometric characteristics. Classification data acquired from remote sensing images have a wide variety of applications. In particular, analysis of remote sensing images have special applications in the classification of various types of vegetation. Results obtained from classification studies of a particular area or region serve towards a greater understanding of what parameters (ecological, temporal, etc.) affect the region being analyzed. In this paper, we make a distinction between both types of classification approaches although, focus is given to the unsupervised classification method using 1987 Thematic Mapped (TM) images of Kennedy Space Center.
Classification of high-resolution multi-swath hyperspectral data using Landsat 8 surface reflectance data as a calibration target and a novel histogram based unsupervised classification technique to determine natural classes from biophysically relevant fit parameters

NASA Astrophysics Data System (ADS)

McCann, C.; Repasky, K. S.; Morin, M.; Lawrence, R. L.; Powell, S. L.

2016-12-01

Compact, cost-effective, flight-based hyperspectral imaging systems can provide scientifically relevant data over large areas for a variety of applications such as ecosystem studies, precision agriculture, and land management. To fully realize this capability, unsupervised classification techniques based on radiometrically-calibrated data that cluster based on biophysical similarity rather than simply spectral similarity are needed. An automated technique to produce high-resolution, large-area, radiometrically-calibrated hyperspectral data sets based on the Landsat surface reflectance data product as a calibration target was developed and applied to three subsequent years of data covering approximately 1850 hectares. The radiometrically-calibrated data allows inter-comparison of the temporal series. Advantages of the radiometric calibration technique include the need for minimal site access, no ancillary instrumentation, and automated processing. Fitting the reflectance spectra of each pixel using a set of biophysically relevant basis functions reduces the data from 80 spectral bands to 9 parameters providing noise reduction and data compression. Examination of histograms of these parameters allows for determination of natural splitting into biophysical similar clusters. This method creates clusters that are similar in terms of biophysical parameters, not simply spectral proximity. Furthermore, this method can be applied to other data sets, such as urban scenes, by developing other physically meaningful basis functions. The ability to use hyperspectral imaging for a variety of important applications requires the development of data processing techniques that can be automated. The radiometric-calibration combined with the histogram based unsupervised classification technique presented here provide one potential avenue for managing big-data associated with hyperspectral imaging.
Automated age-related macular degeneration classification in OCT using unsupervised feature learning

NASA Astrophysics Data System (ADS)

Venhuizen, Freerk G.; van Ginneken, Bram; Bloemen, Bart; van Grinsven, Mark J. J. P.; Philipsen, Rick; Hoyng, Carel; Theelen, Thomas; Sánchez, Clara I.

2015-03-01

Age-related Macular Degeneration (AMD) is a common eye disorder with high prevalence in elderly people. The disease mainly affects the central part of the retina, and could ultimately lead to permanent vision loss. Optical Coherence Tomography (OCT) is becoming the standard imaging modality in diagnosis of AMD and the assessment of its progression. However, the evaluation of the obtained volumetric scan is time consuming, expensive and the signs of early AMD are easy to miss. In this paper we propose a classification method to automatically distinguish AMD patients from healthy subjects with high accuracy. The method is based on an unsupervised feature learning approach, and processes the complete image without the need for an accurate pre-segmentation of the retina. The method can be divided in two steps: an unsupervised clustering stage that extracts a set of small descriptive image patches from the training data, and a supervised training stage that uses these patches to create a patch occurrence histogram for every image on which a random forest classifier is trained. Experiments using 384 volume scans show that the proposed method is capable of identifying AMD patients with high accuracy, obtaining an area under the Receiver Operating Curve of 0:984. Our method allows for a quick and reliable assessment of the presence of AMD pathology in OCT volume scans without the need for accurate layer segmentation algorithms.
Application of classification methods for mapping Mercury's surface composition: analysis on Rudaki's Area

NASA Astrophysics Data System (ADS)

Zambon, F.; De Sanctis, M. C.; Capaccioni, F.; Filacchione, G.; Carli, C.; Ammanito, E.; Friggeri, A.

2011-10-01

During the first two MESSENGER flybys (14th January 2008 and 6th October 2008) the Mercury Dual Imaging System (MDIS) has extended the coverage of the Mercury surface, obtained by Mariner 10 and now we have images of about 90% of the Mercury surface [1]. MDIS is equipped with a Narrow Angle Camera (NAC) and a Wide Angle Camera (WAC). The NAC uses an off-axis reflective design with a 1.5° field of view (FOV) centered at 747 nm. The WAC has a re- fractive design with a 10.5° FOV and 12-position filters that cover a 395-1040 nm spectral range [2]. The color images can be used to infer information on the surface composition and classification meth- ods are an interesting technique for multispectral image analysis which can be applied to the study of the planetary surfaces. Classification methods are based on clustering algorithms and they can be divided in two categories: unsupervised and supervised. The unsupervised classifiers do not require the analyst feedback, and the algorithm automatically organizes pixels values into classes. In the supervised method, instead, the analyst must choose the "training area" that define the pixels value of a given class [3]. Here we will describe the classification in different compositional units of the region near the Rudaki Crater on Mercury.
Unsupervised and self-mapping category formation and semantic object recognition for mobile robot vision used in an actual environment

NASA Astrophysics Data System (ADS)

Madokoro, H.; Tsukada, M.; Sato, K.

2013-07-01

This paper presents an unsupervised learning-based object category formation and recognition method for mobile robot vision. Our method has the following features: detection of feature points and description of features using a scale-invariant feature transform (SIFT), selection of target feature points using one class support vector machines (OC-SVMs), generation of visual words using self-organizing maps (SOMs), formation of labels using adaptive resonance theory 2 (ART-2), and creation and classification of categories on a category map of counter propagation networks (CPNs) for visualizing spatial relations between categories. Classification results of dynamic images using time-series images obtained using two different-size robots and according to movements respectively demonstrate that our method can visualize spatial relations of categories while maintaining time-series characteristics. Moreover, we emphasize the effectiveness of our method for category formation of appearance changes of objects.
Unsupervised identification of cone photoreceptors in non-confocal adaptive optics scanning light ophthalmoscope images.

PubMed

Bergeles, Christos; Dubis, Adam M; Davidson, Benjamin; Kasilian, Melissa; Kalitzeos, Angelos; Carroll, Joseph; Dubra, Alfredo; Michaelides, Michel; Ourselin, Sebastien

2017-06-01

Precise measurements of photoreceptor numerosity and spatial arrangement are promising biomarkers for the early detection of retinal pathologies and may be valuable in the evaluation of retinal therapies. Adaptive optics scanning light ophthalmoscopy (AOSLO) is a method of imaging that corrects for aberrations of the eye to acquire high-resolution images that reveal the photoreceptor mosaic. These images are typically graded manually by experienced observers, obviating the robust, large-scale use of the technology. This paper addresses unsupervised automated detection of cones in non-confocal, split-detection AOSLO images. Our algorithm leverages the appearance of split-detection images to create a cone model that is used for classification. Results show that it compares favorably to the state-of-the-art, both for images of healthy retinas and for images from patients affected by Stargardt disease. The algorithm presented also compares well to manual annotation while excelling in speed.
Automated glioblastoma segmentation based on a multiparametric structured unsupervised classification.

PubMed

Juan-Albarracín, Javier; Fuster-Garcia, Elies; Manjón, José V; Robles, Montserrat; Aparici, F; Martí-Bonmatí, L; García-Gómez, Juan M

2015-01-01

Automatic brain tumour segmentation has become a key component for the future of brain tumour treatment. Currently, most of brain tumour segmentation approaches arise from the supervised learning standpoint, which requires a labelled training dataset from which to infer the models of the classes. The performance of these models is directly determined by the size and quality of the training corpus, whose retrieval becomes a tedious and time-consuming task. On the other hand, unsupervised approaches avoid these limitations but often do not reach comparable results than the supervised methods. In this sense, we propose an automated unsupervised method for brain tumour segmentation based on anatomical Magnetic Resonance (MR) images. Four unsupervised classification algorithms, grouped by their structured or non-structured condition, were evaluated within our pipeline. Considering the non-structured algorithms, we evaluated K-means, Fuzzy K-means and Gaussian Mixture Model (GMM), whereas as structured classification algorithms we evaluated Gaussian Hidden Markov Random Field (GHMRF). An automated postprocess based on a statistical approach supported by tissue probability maps is proposed to automatically identify the tumour classes after the segmentations. We evaluated our brain tumour segmentation method with the public BRAin Tumor Segmentation (BRATS) 2013 Test and Leaderboard datasets. Our approach based on the GMM model improves the results obtained by most of the supervised methods evaluated with the Leaderboard set and reaches the second position in the ranking. Our variant based on the GHMRF achieves the first position in the Test ranking of the unsupervised approaches and the seventh position in the general Test ranking, which confirms the method as a viable alternative for brain tumour segmentation.
Selective Convolutional Descriptor Aggregation for Fine-Grained Image Retrieval.

PubMed

Wei, Xiu-Shen; Luo, Jian-Hao; Wu, Jianxin; Zhou, Zhi-Hua

2017-06-01

Deep convolutional neural network models pre-trained for the ImageNet classification task have been successfully adopted to tasks in other domains, such as texture description and object proposal generation, but these tasks require annotations for images in the new domain. In this paper, we focus on a novel and challenging task in the pure unsupervised setting: fine-grained image retrieval. Even with image labels, fine-grained images are difficult to classify, letting alone the unsupervised retrieval task. We propose the selective convolutional descriptor aggregation (SCDA) method. The SCDA first localizes the main object in fine-grained images, a step that discards the noisy background and keeps useful deep descriptors. The selected descriptors are then aggregated and the dimensionality is reduced into a short feature vector using the best practices we found. The SCDA is unsupervised, using no image label or bounding box annotation. Experiments on six fine-grained data sets confirm the effectiveness of the SCDA for fine-grained image retrieval. Besides, visualization of the SCDA features shows that they correspond to visual attributes (even subtle ones), which might explain SCDA's high-mean average precision in fine-grained retrieval. Moreover, on general image retrieval data sets, the SCDA achieves comparable retrieval results with the state-of-the-art general image retrieval approaches.
Hyperspectral Image Classification using a Self-Organizing Map

NASA Technical Reports Server (NTRS)

Martinez, P.; Gualtieri, J. A.; Aguilar, P. L.; Perez, R. M.; Linaje, M.; Preciado, J. C.; Plaza, A.

2001-01-01

The use of hyperspectral data to determine the abundance of constituents in a certain portion of the Earth's surface relies on the capability of imaging spectrometers to provide a large amount of information at each pixel of a certain scene. Today, hyperspectral imaging sensors are capable of generating unprecedented volumes of radiometric data. The Airborne Visible/Infrared Imaging Spectrometer (AVIRIS), for example, routinely produces image cubes with 224 spectral bands. This undoubtedly opens a wide range of new possibilities, but the analysis of such a massive amount of information is not an easy task. In fact, most of the existing algorithms devoted to analyzing multispectral images are not applicable in the hyperspectral domain, because of the size and high dimensionality of the images. The application of neural networks to perform unsupervised classification of hyperspectral data has been tested by several authors and also by us in some previous work. We have also focused on analyzing the intrinsic capability of neural networks to parallelize the whole hyperspectral unmixing process. The results shown in this work indicate that neural network models are able to find clusters of closely related hyperspectral signatures, and thus can be used as a powerful tool to achieve the desired classification. The present work discusses the possibility of using a Self Organizing neural network to perform unsupervised classification of hyperspectral images. In sections 3 and 4, the topology of the proposed neural network and the training algorithm are respectively described. Section 5 provides the results we have obtained after applying the proposed methodology to real hyperspectral data, described in section 2. Different parameters in the learning stage have been modified in order to obtain a detailed description of their influence on the final results. Finally, in section 6 we provide the conclusions at which we have arrived.
Unsupervised semantic indoor scene classification for robot vision based on context of features using Gist and HSV-SIFT

NASA Astrophysics Data System (ADS)

Madokoro, H.; Yamanashi, A.; Sato, K.

2013-08-01

This paper presents an unsupervised scene classification method for actualizing semantic recognition of indoor scenes. Background and foreground features are respectively extracted using Gist and color scale-invariant feature transform (SIFT) as feature representations based on context. We used hue, saturation, and value SIFT (HSV-SIFT) because of its simple algorithm with low calculation costs. Our method creates bags of features for voting visual words created from both feature descriptors to a two-dimensional histogram. Moreover, our method generates labels as candidates of categories for time-series images while maintaining stability and plasticity together. Automatic labeling of category maps can be realized using labels created using adaptive resonance theory (ART) as teaching signals for counter propagation networks (CPNs). We evaluated our method for semantic scene classification using KTH's image database for robot localization (KTH-IDOL), which is popularly used for robot localization and navigation. The mean classification accuracies of Gist, gray SIFT, one class support vector machines (OC-SVM), position-invariant robust features (PIRF), and our method are, respectively, 39.7, 58.0, 56.0, 63.6, and 79.4%. The result of our method is 15.8% higher than that of PIRF. Moreover, we applied our method for fine classification using our original mobile robot. We obtained mean classification accuracy of 83.2% for six zones.
Classification and analysis of the Rudaki's Area

NASA Astrophysics Data System (ADS)

Zambon, F.; De sanctis, M.; Capaccioni, F.; Filacchione, G.; Carli, C.; Ammannito, E.; Frigeri, A.

2011-12-01

During the first two MESSENGER flybys the Mercury Dual Imaging System (MDIS) has mapped 90% of the Mercury's surface. An effective way to study the different terrain on planetary surfaces is to apply classification methods. These are based on clustering algorithms and they can be divided in two categories: unsupervised and supervised. The unsupervised classifiers do not require the analyst feedback and the algorithm automatically organizes pixels values into classes. In the supervised method, instead, the analyst must choose the "training area" that define the pixels value of a given class. We applied an unsupervised classifier, ISODATA, to the WAC filter images of the Rudaki's area where several kind of terrain have been identified showing differences in albedo, topography and crater density. ISODATA classifier divides this region in four classes: 1) shadow regions, 2) rough regions, 3) smooth plane, 4) highest reflectance area. ISODATA can not distinguish the high albedo regions from highly reflective illuminated edge of the craters, however the algorithm identify four classes that can be considered different units mainly on the basis of their reflectances at the various wavelengths. Is not possible, instead, to extrapolate compositional information because of the absence of clear spectral features. An additional analysis was made using ISODATA to choose the "training area" for further supervised classifications. These approach would allow, for example, to separate more accurately the edge of the craters from the high reflectance areas and the low reflectance regions from the shadow areas.
Manifold Learning in MR spectroscopy using nonlinear dimensionality reduction and unsupervised clustering.

PubMed

Yang, Guang; Raschke, Felix; Barrick, Thomas R; Howe, Franklyn A

2015-09-01

To investigate whether nonlinear dimensionality reduction improves unsupervised classification of (1) H MRS brain tumor data compared with a linear method. In vivo single-voxel (1) H magnetic resonance spectroscopy (55 patients) and (1) H magnetic resonance spectroscopy imaging (MRSI) (29 patients) data were acquired from histopathologically diagnosed gliomas. Data reduction using Laplacian eigenmaps (LE) or independent component analysis (ICA) was followed by k-means clustering or agglomerative hierarchical clustering (AHC) for unsupervised learning to assess tumor grade and for tissue type segmentation of MRSI data. An accuracy of 93% in classification of glioma grade II and grade IV, with 100% accuracy in distinguishing tumor and normal spectra, was obtained by LE with unsupervised clustering, but not with the combination of k-means and ICA. With (1) H MRSI data, LE provided a more linear distribution of data for cluster analysis and better cluster stability than ICA. LE combined with k-means or AHC provided 91% accuracy for classifying tumor grade and 100% accuracy for identifying normal tissue voxels. Color-coded visualization of normal brain, tumor core, and infiltration regions was achieved with LE combined with AHC. The LE method is promising for unsupervised clustering to separate brain and tumor tissue with automated color-coding for visualization of (1) H MRSI data after cluster analysis. © 2014 Wiley Periodicals, Inc.
An unsupervised classification technique for multispectral remote sensing data.

NASA Technical Reports Server (NTRS)

Su, M. Y.; Cummings, R. E.

1973-01-01

Description of a two-part clustering technique consisting of (a) a sequential statistical clustering, which is essentially a sequential variance analysis, and (b) a generalized K-means clustering. In this composite clustering technique, the output of (a) is a set of initial clusters which are input to (b) for further improvement by an iterative scheme. This unsupervised composite technique was employed for automatic classification of two sets of remote multispectral earth resource observations. The classification accuracy by the unsupervised technique is found to be comparable to that by traditional supervised maximum-likelihood classification techniques.
Unsupervised classification of earth resources data.

NASA Technical Reports Server (NTRS)

Su, M. Y.; Jayroe, R. R., Jr.; Cummings, R. E.

1972-01-01

A new clustering technique is presented. It consists of two parts: (a) a sequential statistical clustering which is essentially a sequential variance analysis and (b) a generalized K-means clustering. In this composite clustering technique, the output of (a) is a set of initial clusters which are input to (b) for further improvement by an iterative scheme. This unsupervised composite technique was employed for automatic classification of two sets of remote multispectral earth resource observations. The classification accuracy by the unsupervised technique is found to be comparable to that by existing supervised maximum liklihood classification technique.

Automated Glioblastoma Segmentation Based on a Multiparametric Structured Unsupervised Classification

PubMed Central

Juan-Albarracín, Javier; Fuster-Garcia, Elies; Manjón, José V.; Robles, Montserrat; Aparici, F.; Martí-Bonmatí, L.; García-Gómez, Juan M.

2015-01-01

Automatic brain tumour segmentation has become a key component for the future of brain tumour treatment. Currently, most of brain tumour segmentation approaches arise from the supervised learning standpoint, which requires a labelled training dataset from which to infer the models of the classes. The performance of these models is directly determined by the size and quality of the training corpus, whose retrieval becomes a tedious and time-consuming task. On the other hand, unsupervised approaches avoid these limitations but often do not reach comparable results than the supervised methods. In this sense, we propose an automated unsupervised method for brain tumour segmentation based on anatomical Magnetic Resonance (MR) images. Four unsupervised classification algorithms, grouped by their structured or non-structured condition, were evaluated within our pipeline. Considering the non-structured algorithms, we evaluated K-means, Fuzzy K-means and Gaussian Mixture Model (GMM), whereas as structured classification algorithms we evaluated Gaussian Hidden Markov Random Field (GHMRF). An automated postprocess based on a statistical approach supported by tissue probability maps is proposed to automatically identify the tumour classes after the segmentations. We evaluated our brain tumour segmentation method with the public BRAin Tumor Segmentation (BRATS) 2013 Test and Leaderboard datasets. Our approach based on the GMM model improves the results obtained by most of the supervised methods evaluated with the Leaderboard set and reaches the second position in the ranking. Our variant based on the GHMRF achieves the first position in the Test ranking of the unsupervised approaches and the seventh position in the general Test ranking, which confirms the method as a viable alternative for brain tumour segmentation. PMID:25978453
Evaluating unsupervised methods to size and classify suspended particles using digital in-line holography

USGS Publications Warehouse

Davies, Emlyn J.; Buscombe, Daniel D.; Graham, George W.; Nimmo-Smith, W. Alex M.

2015-01-01

Substantial information can be gained from digital in-line holography of marine particles, eliminating depth-of-field and focusing errors associated with standard lens-based imaging methods. However, for the technique to reach its full potential in oceanographic research, fully unsupervised (automated) methods are required for focusing, segmentation, sizing and classification of particles. These computational challenges are the subject of this paper, in which we draw upon data collected using a variety of holographic systems developed at Plymouth University, UK, from a significant range of particle types, sizes and shapes. A new method for noise reduction in reconstructed planes is found to be successful in aiding particle segmentation and sizing. The performance of an automated routine for deriving particle characteristics (and subsequent size distributions) is evaluated against equivalent size metrics obtained by a trained operative measuring grain axes on screen. The unsupervised method is found to be reliable, despite some errors resulting from over-segmentation of particles. A simple unsupervised particle classification system is developed, and is capable of successfully differentiating sand grains, bubbles and diatoms from within the surf-zone. Avoiding miscounting bubbles and biological particles as sand grains enables more accurate estimates of sand concentrations, and is especially important in deployments of particle monitoring instrumentation in aerated water. Perhaps the greatest potential for further development in the computational aspects of particle holography is in the area of unsupervised particle classification. The simple method proposed here provides a foundation upon which further development could lead to reliable identification of more complex particle populations, such as those containing phytoplankton, zooplankton, flocculated cohesive sediments and oil droplets.
VHR satellite multitemporal data to extract cultural landscape changes in the roman site of Grumentum

NASA Astrophysics Data System (ADS)

masini, nicola; Lasaponara, Rosa

2013-04-01

The papers deals with the use of VHR satellite multitemporal data set to extract cultural landscape changes in the roman site of Grumentum Grumentum is an ancient town, 50 km south of Potenza, located near the roman road of Via Herculea which connected the Venusia, in the north est of Basilicata, with Heraclea in the Ionian coast. The first settlement date back to the 6th century BC. It was resettled by the Romans in the 3rd century BC. Its urban fabric which evidences a long history from the Republican age to late Antiquity (III BC-V AD) is composed of the typical urban pattern of cardi and decumani. Its excavated ruins include a large amphitheatre, a theatre, the thermae, the Forum and some temples. There are many techniques nowadays available to capture and record differences in two or more images. In this paper we focus and apply the two main approaches which can be distinguished into : (i) unsupervised and (ii) supervised change detection methods. Unsupervised change detection methods are generally based on the transformation of the two multispectral images in to a single band or multiband image which are further analyzed to identify changes Unsupervised change detection techniques are generally based on three basic steps (i) the preprocessing step, (ii) a pixel-by-pixel comparison is performed, (iii). Identification of changes according to the magnitude an direction (positive /negative). Unsupervised change detection are generally based on the transformation of the two multispectral images into a single band or multiband image which are further analyzed to identify changes. Than the separation between changed and unchanged classes is obtained from the magnitude of the resulting spectral change vectors by means of empirical or theoretical well founded approaches Supervised change detection methods are generally based on supervised classification methods, which require the availability of a suitable training set for the learning process of the classifiers. Unsupervised change detection techniques are generally based on three basic steps (i) the preprocessing step, (ii) supervised classification is performed on the single dates or on the map obtained as the difference of two dates, (iii). Identification of changes according to the magnitude an direction (positive /negative). Supervised change detection are generally based on supervised classification methods, which require the availability of a suitable training set for the learning process of the classifiers, therefore these algorithms require a preliminary knowledge necessary: (i) to generate representative parameters for each class of interest; and (ii) to carry out the training stage Advantages and disadvantages of the supervised and unsupervised approaches are discuss. Finally results from the the satellite multitemporal dataset was also integrated with aerial photos from historical archive in order to expand the time window of the investigation and capture landscape changes occurred from the Agrarian Reform, in the 50s, up today.
Rational Variety Mapping for Contrast-Enhanced Nonlinear Unsupervised Segmentation of Multispectral Images of Unstained Specimen

PubMed Central

Kopriva, Ivica; Hadžija, Mirko; Popović Hadžija, Marijana; Korolija, Marina; Cichocki, Andrzej

2011-01-01

A methodology is proposed for nonlinear contrast-enhanced unsupervised segmentation of multispectral (color) microscopy images of principally unstained specimens. The methodology exploits spectral diversity and spatial sparseness to find anatomical differences between materials (cells, nuclei, and background) present in the image. It consists of rth-order rational variety mapping (RVM) followed by matrix/tensor factorization. Sparseness constraint implies duality between nonlinear unsupervised segmentation and multiclass pattern assignment problems. Classes not linearly separable in the original input space become separable with high probability in the higher-dimensional mapped space. Hence, RVM mapping has two advantages: it takes implicitly into account nonlinearities present in the image (ie, they are not required to be known) and it increases spectral diversity (ie, contrast) between materials, due to increased dimensionality of the mapped space. This is expected to improve performance of systems for automated classification and analysis of microscopic histopathological images. The methodology was validated using RVM of the second and third orders of the experimental multispectral microscopy images of unstained sciatic nerve fibers (nervus ischiadicus) and of unstained white pulp in the spleen tissue, compared with a manually defined ground truth labeled by two trained pathophysiologists. The methodology can also be useful for additional contrast enhancement of images of stained specimens. PMID:21708116
Comparison of unsupervised classification methods for brain tumor segmentation using multi-parametric MRI.

PubMed

Sauwen, N; Acou, M; Van Cauter, S; Sima, D M; Veraart, J; Maes, F; Himmelreich, U; Achten, E; Van Huffel, S

2016-01-01

Tumor segmentation is a particularly challenging task in high-grade gliomas (HGGs), as they are among the most heterogeneous tumors in oncology. An accurate delineation of the lesion and its main subcomponents contributes to optimal treatment planning, prognosis and follow-up. Conventional MRI (cMRI) is the imaging modality of choice for manual segmentation, and is also considered in the vast majority of automated segmentation studies. Advanced MRI modalities such as perfusion-weighted imaging (PWI), diffusion-weighted imaging (DWI) and magnetic resonance spectroscopic imaging (MRSI) have already shown their added value in tumor tissue characterization, hence there have been recent suggestions of combining different MRI modalities into a multi-parametric MRI (MP-MRI) approach for brain tumor segmentation. In this paper, we compare the performance of several unsupervised classification methods for HGG segmentation based on MP-MRI data including cMRI, DWI, MRSI and PWI. Two independent MP-MRI datasets with a different acquisition protocol were available from different hospitals. We demonstrate that a hierarchical non-negative matrix factorization variant which was previously introduced for MP-MRI tumor segmentation gives the best performance in terms of mean Dice-scores for the pathologic tissue classes on both datasets.
DISCRIMINATION OF GRANITOIDS AND MINERALIZED GRANITOIDS IN THE MIDYAN REGION, NORTHWESTERN ARABIAN SHIELD, SAUDI ARABIA, BY LANDSAT MSS DATA-ANALYSIS.

USGS Publications Warehouse

Davis, Philip A.; Grolier, Maurice J.

1984-01-01

Landsat multispectral scanner (MSS) band and band-ratio databases of two scenes covering the Midyan region of northwestern Saudi Arabia were examined quantitatively and qualitatively to determine which databases best discriminate the geologic units of this semi-arid and arid region. Unsupervised, linear-discriminant cluster-analysis was performed on these two band-ratio combinations and on the MSS bands for both scenes. The results for granitoid-rock discrimination indicated that the classification images using the MSS bands are superior to the band-ratio classification images for two reasons, discussed in the paper. Yet, the effects of topography and material type (including desert varnish) on the MSS-band data produced ambiguities in the MSS-band classification results. However, these ambiguities were clarified by using a simulated natural-color image in conjunction with the MSS-band classification image.
Design of partially supervised classifiers for multispectral image data

NASA Technical Reports Server (NTRS)

Jeon, Byeungwoo; Landgrebe, David

1993-01-01

A partially supervised classification problem is addressed, especially when the class definition and corresponding training samples are provided a priori only for just one particular class. In practical applications of pattern classification techniques, a frequently observed characteristic is the heavy, often nearly impossible requirements on representative prior statistical class characteristics of all classes in a given data set. Considering the effort in both time and man-power required to have a well-defined, exhaustive list of classes with a corresponding representative set of training samples, this 'partially' supervised capability would be very desirable, assuming adequate classifier performance can be obtained. Two different classification algorithms are developed to achieve simplicity in classifier design by reducing the requirement of prior statistical information without sacrificing significant classifying capability. The first one is based on optimal significance testing, where the optimal acceptance probability is estimated directly from the data set. In the second approach, the partially supervised classification is considered as a problem of unsupervised clustering with initially one known cluster or class. A weighted unsupervised clustering procedure is developed to automatically define other classes and estimate their class statistics. The operational simplicity thus realized should make these partially supervised classification schemes very viable tools in pattern classification.
A semi-supervised classification algorithm using the TAD-derived background as training data

NASA Astrophysics Data System (ADS)

Fan, Lei; Ambeau, Brittany; Messinger, David W.

2013-05-01

In general, spectral image classification algorithms fall into one of two categories: supervised and unsupervised. In unsupervised approaches, the algorithm automatically identifies clusters in the data without a priori information about those clusters (except perhaps the expected number of them). Supervised approaches require an analyst to identify training data to learn the characteristics of the clusters such that they can then classify all other pixels into one of the pre-defined groups. The classification algorithm presented here is a semi-supervised approach based on the Topological Anomaly Detection (TAD) algorithm. The TAD algorithm defines background components based on a mutual k-Nearest Neighbor graph model of the data, along with a spectral connected components analysis. Here, the largest components produced by TAD are used as regions of interest (ROI's),or training data for a supervised classification scheme. By combining those ROI's with a Gaussian Maximum Likelihood (GML) or a Minimum Distance to the Mean (MDM) algorithm, we are able to achieve a semi supervised classification method. We test this classification algorithm against data collected by the HyMAP sensor over the Cooke City, MT area and University of Pavia scene.
Spatio-spectral classification of hyperspectral images for brain cancer detection during surgical operations.

PubMed

Fabelo, Himar; Ortega, Samuel; Ravi, Daniele; Kiran, B Ravi; Sosa, Coralia; Bulters, Diederik; Callicó, Gustavo M; Bulstrode, Harry; Szolna, Adam; Piñeiro, Juan F; Kabwama, Silvester; Madroñal, Daniel; Lazcano, Raquel; J-O'Shanahan, Aruma; Bisshopp, Sara; Hernández, María; Báez, Abelardo; Yang, Guang-Zhong; Stanciulescu, Bogdan; Salvador, Rubén; Juárez, Eduardo; Sarmiento, Roberto

2018-01-01

Surgery for brain cancer is a major problem in neurosurgery. The diffuse infiltration into the surrounding normal brain by these tumors makes their accurate identification by the naked eye difficult. Since surgery is the common treatment for brain cancer, an accurate radical resection of the tumor leads to improved survival rates for patients. However, the identification of the tumor boundaries during surgery is challenging. Hyperspectral imaging is a non-contact, non-ionizing and non-invasive technique suitable for medical diagnosis. This study presents the development of a novel classification method taking into account the spatial and spectral characteristics of the hyperspectral images to help neurosurgeons to accurately determine the tumor boundaries in surgical-time during the resection, avoiding excessive excision of normal tissue or unintentionally leaving residual tumor. The algorithm proposed in this study to approach an efficient solution consists of a hybrid framework that combines both supervised and unsupervised machine learning methods. Firstly, a supervised pixel-wise classification using a Support Vector Machine classifier is performed. The generated classification map is spatially homogenized using a one-band representation of the HS cube, employing the Fixed Reference t-Stochastic Neighbors Embedding dimensional reduction algorithm, and performing a K-Nearest Neighbors filtering. The information generated by the supervised stage is combined with a segmentation map obtained via unsupervised clustering employing a Hierarchical K-Means algorithm. The fusion is performed using a majority voting approach that associates each cluster with a certain class. To evaluate the proposed approach, five hyperspectral images of surface of the brain affected by glioblastoma tumor in vivo from five different patients have been used. The final classification maps obtained have been analyzed and validated by specialists. These preliminary results are promising, obtaining an accurate delineation of the tumor area.
Spatio-spectral classification of hyperspectral images for brain cancer detection during surgical operations

PubMed Central

Kabwama, Silvester; Madroñal, Daniel; Lazcano, Raquel; J-O’Shanahan, Aruma; Bisshopp, Sara; Hernández, María; Báez, Abelardo; Yang, Guang-Zhong; Stanciulescu, Bogdan; Salvador, Rubén; Juárez, Eduardo; Sarmiento, Roberto

2018-01-01

Surgery for brain cancer is a major problem in neurosurgery. The diffuse infiltration into the surrounding normal brain by these tumors makes their accurate identification by the naked eye difficult. Since surgery is the common treatment for brain cancer, an accurate radical resection of the tumor leads to improved survival rates for patients. However, the identification of the tumor boundaries during surgery is challenging. Hyperspectral imaging is a non-contact, non-ionizing and non-invasive technique suitable for medical diagnosis. This study presents the development of a novel classification method taking into account the spatial and spectral characteristics of the hyperspectral images to help neurosurgeons to accurately determine the tumor boundaries in surgical-time during the resection, avoiding excessive excision of normal tissue or unintentionally leaving residual tumor. The algorithm proposed in this study to approach an efficient solution consists of a hybrid framework that combines both supervised and unsupervised machine learning methods. Firstly, a supervised pixel-wise classification using a Support Vector Machine classifier is performed. The generated classification map is spatially homogenized using a one-band representation of the HS cube, employing the Fixed Reference t-Stochastic Neighbors Embedding dimensional reduction algorithm, and performing a K-Nearest Neighbors filtering. The information generated by the supervised stage is combined with a segmentation map obtained via unsupervised clustering employing a Hierarchical K-Means algorithm. The fusion is performed using a majority voting approach that associates each cluster with a certain class. To evaluate the proposed approach, five hyperspectral images of surface of the brain affected by glioblastoma tumor in vivo from five different patients have been used. The final classification maps obtained have been analyzed and validated by specialists. These preliminary results are promising, obtaining an accurate delineation of the tumor area. PMID:29554126
Unsupervised Feature Learning With Winner-Takes-All Based STDP

PubMed Central

Ferré, Paul; Mamalet, Franck; Thorpe, Simon J.

2018-01-01

We present a novel strategy for unsupervised feature learning in image applications inspired by the Spike-Timing-Dependent-Plasticity (STDP) biological learning rule. We show equivalence between rank order coding Leaky-Integrate-and-Fire neurons and ReLU artificial neurons when applied to non-temporal data. We apply this to images using rank-order coding, which allows us to perform a full network simulation with a single feed-forward pass using GPU hardware. Next we introduce a binary STDP learning rule compatible with training on batches of images. Two mechanisms to stabilize the training are also presented : a Winner-Takes-All (WTA) framework which selects the most relevant patches to learn from along the spatial dimensions, and a simple feature-wise normalization as homeostatic process. This learning process allows us to train multi-layer architectures of convolutional sparse features. We apply our method to extract features from the MNIST, ETH80, CIFAR-10, and STL-10 datasets and show that these features are relevant for classification. We finally compare these results with several other state of the art unsupervised learning methods. PMID:29674961
Unsupervised classification of major depression using functional connectivity MRI.

PubMed

Zeng, Ling-Li; Shen, Hui; Liu, Li; Hu, Dewen

2014-04-01

The current diagnosis of psychiatric disorders including major depressive disorder based largely on self-reported symptoms and clinical signs may be prone to patients' behaviors and psychiatrists' bias. This study aims at developing an unsupervised machine learning approach for the accurate identification of major depression based on single resting-state functional magnetic resonance imaging scans in the absence of clinical information. Twenty-four medication-naive patients with major depression and 29 demographically similar healthy individuals underwent resting-state functional magnetic resonance imaging. We first clustered the voxels within the perigenual cingulate cortex into two subregions, a subgenual region and a pregenual region, according to their distinct resting-state functional connectivity patterns and showed that a maximum margin clustering-based unsupervised machine learning approach extracted sufficient information from the subgenual cingulate functional connectivity map to differentiate depressed patients from healthy controls with a group-level clustering consistency of 92.5% and an individual-level classification consistency of 92.5%. It was also revealed that the subgenual cingulate functional connectivity network with the highest discriminative power primarily included the ventrolateral and ventromedial prefrontal cortex, superior temporal gyri and limbic areas, indicating that these connections may play critical roles in the pathophysiology of major depression. The current study suggests that subgenual cingulate functional connectivity network signatures may provide promising objective biomarkers for the diagnosis of major depression and that maximum margin clustering-based unsupervised machine learning approaches may have the potential to inform clinical practice and aid in research on psychiatric disorders. Copyright © 2013 Wiley Periodicals, Inc.
Unsupervised change detection in a particular vegetation land cover type using spectral angle mapper

NASA Astrophysics Data System (ADS)

Renza, Diego; Martinez, Estibaliz; Molina, Iñigo; Ballesteros L., Dora M.

2017-04-01

This paper presents a new unsupervised change detection methodology for multispectral images applied to specific land covers. The proposed method involves comparing each image against a reference spectrum, where the reference spectrum is obtained from the spectral signature of the type of coverage you want to detect. In this case the method has been tested using multispectral images (SPOT5) of the community of Madrid (Spain), and multispectral images (Quickbird) of an area over Indonesia that was impacted by the December 26, 2004 tsunami; here, the tests have focused on the detection of changes in vegetation. The image comparison is obtained by applying Spectral Angle Mapper between the reference spectrum and each multitemporal image. Then, a threshold to produce a single image of change is applied, which corresponds to the vegetation zones. The results for each multitemporal image are combined through an exclusive or (XOR) operation that selects vegetation zones that have changed over time. Finally, the derived results were compared against a supervised method based on classification with the Support Vector Machine. Furthermore, the NDVI-differencing and the Spectral Angle Mapper techniques were selected as unsupervised methods for comparison purposes. The main novelty of the method consists in the detection of changes in a specific land cover type (vegetation), therefore, for comparison purposes, the best scenario is to compare it with methods that aim to detect changes in a specific land cover type (vegetation). This is the main reason to select NDVI-based method and the post-classification method (SVM implemented in a standard software tool). To evaluate the improvements using a reference spectrum vector, the results are compared with the basic-SAM method. In SPOT5 image, the overall accuracy was 99.36% and the κ index was 90.11%; in Quickbird image, the overall accuracy was 97.5% and the κ index was 82.16%. Finally, the precision results of the method are comparable to those of a supervised method, supported by low detection of false positives and false negatives, along with a high overall accuracy and a high kappa index. On the other hand, the execution times were comparable to those of unsupervised methods of low computational load.
An Efficient Optimization Method for Solving Unsupervised Data Classification Problems.

PubMed

Shabanzadeh, Parvaneh; Yusof, Rubiyah

2015-01-01

Unsupervised data classification (or clustering) analysis is one of the most useful tools and a descriptive task in data mining that seeks to classify homogeneous groups of objects based on similarity and is used in many medical disciplines and various applications. In general, there is no single algorithm that is suitable for all types of data, conditions, and applications. Each algorithm has its own advantages, limitations, and deficiencies. Hence, research for novel and effective approaches for unsupervised data classification is still active. In this paper a heuristic algorithm, Biogeography-Based Optimization (BBO) algorithm, was adapted for data clustering problems by modifying the main operators of BBO algorithm, which is inspired from the natural biogeography distribution of different species. Similar to other population-based algorithms, BBO algorithm starts with an initial population of candidate solutions to an optimization problem and an objective function that is calculated for them. To evaluate the performance of the proposed algorithm assessment was carried on six medical and real life datasets and was compared with eight well known and recent unsupervised data classification algorithms. Numerical results demonstrate that the proposed evolutionary optimization algorithm is efficient for unsupervised data classification.
Evaluating unsupervised and supervised image classification methods for mapping cotton root rot

USDA-ARS?s Scientific Manuscript database

Cotton root rot, caused by the soilborne fungus Phymatotrichopsis omnivora, is one of the most destructive plant diseases occurring throughout the southwestern United States. This disease has plagued the cotton industry for over a century, but effective practices for its control are still lacking. R...
Hierarchical classification method and its application in shape representation

NASA Astrophysics Data System (ADS)

Ireton, M. A.; Oakley, John P.; Xydeas, Costas S.

1992-04-01

In this paper we describe a technique for performing shaped-based content retrieval of images from a large database. In order to be able to formulate such user-generated queries about visual objects, we have developed an hierarchical classification technique. This hierarchical classification technique enables similarity matching between objects, with the position in the hierarchy signifying the level of generality to be used in the query. The classification technique is unsupervised, robust, and general; it can be applied to any suitable parameter set. To establish the potential of this classifier for aiding visual querying, we have applied it to the classification of the 2-D outlines of leaves.
Korean coastal water depth/sediment and land cover mapping (1:25,000) by computer analysis of LANDSAT imagery

NASA Technical Reports Server (NTRS)

Park, K. Y.; Miller, L. D.

1978-01-01

Computer analysis was applied to single date LANDSAT MSS imagery of a sample coastal area near Seoul, Korea equivalent to a 1:50,000 topographic map. Supervised image processing yielded a test classification map from this sample image containing 12 classes: 5 water depth/sediment classes, 2 shoreline/tidal classes, and 5 coastal land cover classes at a scale of 1:25,000 and with a training set accuracy of 76%. Unsupervised image classification was applied to a subportion of the site analyzed and produced classification maps comparable in results in a spatial sense. The results of this test indicated that it is feasible to produce such quantitative maps for detailed study of dynamic coastal processes given a LANDSAT image data base at sufficiently frequent time intervals.
Age and gender classification in the wild with unsupervised feature learning

NASA Astrophysics Data System (ADS)

Wan, Lihong; Huo, Hong; Fang, Tao

2017-03-01

Inspired by unsupervised feature learning (UFL) within the self-taught learning framework, we propose a method based on UFL, convolution representation, and part-based dimensionality reduction to handle facial age and gender classification, which are two challenging problems under unconstrained circumstances. First, UFL is introduced to learn selective receptive fields (filters) automatically by applying whitening transformation and spherical k-means on random patches collected from unlabeled data. The learning process is fast and has no hyperparameters to tune. Then, the input image is convolved with these filters to obtain filtering responses on which local contrast normalization is applied. Average pooling and feature concatenation are then used to form global face representation. Finally, linear discriminant analysis with part-based strategy is presented to reduce the dimensions of the global representation and to improve classification performances further. Experiments on three challenging databases, namely, Labeled faces in the wild, Gallagher group photos, and Adience, demonstrate the effectiveness of the proposed method relative to that of state-of-the-art approaches.
Land cover classification in multispectral imagery using clustering of sparse approximations over learned feature dictionaries

DOE PAGES

Moody, Daniela I.; Brumby, Steven P.; Rowland, Joel C.; ...

2014-12-09

We present results from an ongoing effort to extend neuromimetic machine vision algorithms to multispectral data using adaptive signal processing combined with compressive sensing and machine learning techniques. Our goal is to develop a robust classification methodology that will allow for automated discretization of the landscape into distinct units based on attributes such as vegetation, surface hydrological properties, and topographic/geomorphic characteristics. We use a Hebbian learning rule to build spectral-textural dictionaries that are tailored for classification. We learn our dictionaries from millions of overlapping multispectral image patches and then use a pursuit search to generate classification features. Land cover labelsmore » are automatically generated using unsupervised clustering of sparse approximations (CoSA). We demonstrate our method on multispectral WorldView-2 data from a coastal plain ecosystem in Barrow, Alaska. We explore learning from both raw multispectral imagery and normalized band difference indices. We explore a quantitative metric to evaluate the spectral properties of the clusters in order to potentially aid in assigning land cover categories to the cluster labels. In this study, our results suggest CoSA is a promising approach to unsupervised land cover classification in high-resolution satellite imagery.« less
An Unsupervised kNN Method to Systematically Detect Changes in Protein Localization in High-Throughput Microscopy Images.

PubMed

Lu, Alex Xijie; Moses, Alan M

2016-01-01

Despite the importance of characterizing genes that exhibit subcellular localization changes between conditions in proteome-wide imaging experiments, many recent studies still rely upon manual evaluation to assess the results of high-throughput imaging experiments. We describe and demonstrate an unsupervised k-nearest neighbours method for the detection of localization changes. Compared to previous classification-based supervised change detection methods, our method is much simpler and faster, and operates directly on the feature space to overcome limitations in needing to manually curate training sets that may not generalize well between screens. In addition, the output of our method is flexible in its utility, generating both a quantitatively ranked list of localization changes that permit user-defined cut-offs, and a vector for each gene describing feature-wise direction and magnitude of localization changes. We demonstrate that our method is effective at the detection of localization changes using the Δrpd3 perturbation in Saccharomyces cerevisiae, where we capture 71.4% of previously known changes within the top 10% of ranked genes, and find at least four new localization changes within the top 1% of ranked genes. The results of our analysis indicate that simple unsupervised methods may be able to identify localization changes in images without laborious manual image labelling steps.

Classification by Using Multispectral Point Cloud Data

NASA Astrophysics Data System (ADS)

Liao, C. T.; Huang, H. H.

2012-07-01

Remote sensing images are generally recorded in two-dimensional format containing multispectral information. Also, the semantic information is clearly visualized, which ground features can be better recognized and classified via supervised or unsupervised classification methods easily. Nevertheless, the shortcomings of multispectral images are highly depending on light conditions, and classification results lack of three-dimensional semantic information. On the other hand, LiDAR has become a main technology for acquiring high accuracy point cloud data. The advantages of LiDAR are high data acquisition rate, independent of light conditions and can directly produce three-dimensional coordinates. However, comparing with multispectral images, the disadvantage is multispectral information shortage, which remains a challenge in ground feature classification through massive point cloud data. Consequently, by combining the advantages of both LiDAR and multispectral images, point cloud data with three-dimensional coordinates and multispectral information can produce a integrate solution for point cloud classification. Therefore, this research acquires visible light and near infrared images, via close range photogrammetry, by matching images automatically through free online service for multispectral point cloud generation. Then, one can use three-dimensional affine coordinate transformation to compare the data increment. At last, the given threshold of height and color information is set as threshold in classification.
Epithelium-Stroma Classification via Convolutional Neural Networks and Unsupervised Domain Adaptation in Histopathological Images.

PubMed

Huang, Yue; Zheng, Han; Liu, Chi; Ding, Xinghao; Rohde, Gustavo K

2017-11-01

Epithelium-stroma classification is a necessary preprocessing step in histopathological image analysis. Current deep learning based recognition methods for histology data require collection of large volumes of labeled data in order to train a new neural network when there are changes to the image acquisition procedure. However, it is extremely expensive for pathologists to manually label sufficient volumes of data for each pathology study in a professional manner, which results in limitations in real-world applications. A very simple but effective deep learning method, that introduces the concept of unsupervised domain adaptation to a simple convolutional neural network (CNN), has been proposed in this paper. Inspired by transfer learning, our paper assumes that the training data and testing data follow different distributions, and there is an adaptation operation to more accurately estimate the kernels in CNN in feature extraction, in order to enhance performance by transferring knowledge from labeled data in source domain to unlabeled data in target domain. The model has been evaluated using three independent public epithelium-stroma datasets by cross-dataset validations. The experimental results demonstrate that for epithelium-stroma classification, the proposed framework outperforms the state-of-the-art deep neural network model, and it also achieves better performance than other existing deep domain adaptation methods. The proposed model can be considered to be a better option for real-world applications in histopathological image analysis, since there is no longer a requirement for large-scale labeled data in each specified domain.
Rational variety mapping for contrast-enhanced nonlinear unsupervised segmentation of multispectral images of unstained specimen.

PubMed

Kopriva, Ivica; Hadžija, Mirko; Popović Hadžija, Marijana; Korolija, Marina; Cichocki, Andrzej

2011-08-01

A methodology is proposed for nonlinear contrast-enhanced unsupervised segmentation of multispectral (color) microscopy images of principally unstained specimens. The methodology exploits spectral diversity and spatial sparseness to find anatomical differences between materials (cells, nuclei, and background) present in the image. It consists of rth-order rational variety mapping (RVM) followed by matrix/tensor factorization. Sparseness constraint implies duality between nonlinear unsupervised segmentation and multiclass pattern assignment problems. Classes not linearly separable in the original input space become separable with high probability in the higher-dimensional mapped space. Hence, RVM mapping has two advantages: it takes implicitly into account nonlinearities present in the image (ie, they are not required to be known) and it increases spectral diversity (ie, contrast) between materials, due to increased dimensionality of the mapped space. This is expected to improve performance of systems for automated classification and analysis of microscopic histopathological images. The methodology was validated using RVM of the second and third orders of the experimental multispectral microscopy images of unstained sciatic nerve fibers (nervus ischiadicus) and of unstained white pulp in the spleen tissue, compared with a manually defined ground truth labeled by two trained pathophysiologists. The methodology can also be useful for additional contrast enhancement of images of stained specimens. Copyright © 2011 American Society for Investigative Pathology. Published by Elsevier Inc. All rights reserved.
Detection of land-use and land cover changes in Franklin, Gulf, and Liberty Counties, Florida, with multitemporal landsat thematic mapper images

Treesearch

Shufen Pan; Guiying Li

2007-01-01

Florida Panhandle region has been experiencing rapid land transformation in the recent decades. To quantify land use and land-cover (LULC) changes and other landscape changes in this area, three counties including Franklin, Liberty and Gulf were taken as a case study and an unsupervised classification approach implemented to Landsat TM images acquired from 1985 to 2005...
An unsupervised classification approach for analysis of Landsat data to monitor land reclamation in Belmont county, Ohio

NASA Technical Reports Server (NTRS)

Brumfield, J. O.; Bloemer, H. H. L.; Campbell, W. J.

1981-01-01

Two unsupervised classification procedures for analyzing Landsat data used to monitor land reclamation in a surface mining area in east central Ohio are compared for agreement with data collected from the corresponding locations on the ground. One procedure is based on a traditional unsupervised-clustering/maximum-likelihood algorithm sequence that assumes spectral groupings in the Landsat data in n-dimensional space; the other is based on a nontraditional unsupervised-clustering/canonical-transformation/clustering algorithm sequence that not only assumes spectral groupings in n-dimensional space but also includes an additional feature-extraction technique. It is found that the nontraditional procedure provides an appreciable improvement in spectral groupings and apparently increases the level of accuracy in the classification of land cover categories.
Use of Binary Partition Tree and energy minimization for object-based classification of urban land cover

NASA Astrophysics Data System (ADS)

Li, Mengmeng; Bijker, Wietske; Stein, Alfred

2015-04-01

Two main challenges are faced when classifying urban land cover from very high resolution satellite images: obtaining an optimal image segmentation and distinguishing buildings from other man-made objects. For optimal segmentation, this work proposes a hierarchical representation of an image by means of a Binary Partition Tree (BPT) and an unsupervised evaluation of image segmentations by energy minimization. For building extraction, we apply fuzzy sets to create a fuzzy landscape of shadows which in turn involves a two-step procedure. The first step is a preliminarily image classification at a fine segmentation level to generate vegetation and shadow information. The second step models the directional relationship between building and shadow objects to extract building information at the optimal segmentation level. We conducted the experiments on two datasets of Pléiades images from Wuhan City, China. To demonstrate its performance, the proposed classification is compared at the optimal segmentation level with Maximum Likelihood Classification and Support Vector Machine classification. The results show that the proposed classification produced the highest overall accuracies and kappa coefficients, and the smallest over-classification and under-classification geometric errors. We conclude first that integrating BPT with energy minimization offers an effective means for image segmentation. Second, we conclude that the directional relationship between building and shadow objects represented by a fuzzy landscape is important for building extraction.
The evaluation of alternate methodologies for land cover classification in an urbanizing area

NASA Technical Reports Server (NTRS)

Smekofski, R. M.

1981-01-01

The usefulness of LANDSAT in classifying land cover and in identifying and classifying land use change was investigated using an urbanizing area as the study area. The question of what was the best technique for classification was the primary focus of the study. The many computer-assisted techniques available to analyze LANDSAT data were evaluated. Techniques of statistical training (polygons from CRT, unsupervised clustering, polygons from digitizer and binary masks) were tested with minimum distance to the mean, maximum likelihood and canonical analysis with minimum distance to the mean classifiers. The twelve output images were compared to photointerpreted samples, ground verified samples and a current land use data base. Results indicate that for a reconnaissance inventory, the unsupervised training with canonical analysis-minimum distance classifier is the most efficient. If more detailed ground truth and ground verification is available, the polygons from the digitizer training with the canonical analysis minimum distance is more accurate.
Multilayer Extreme Learning Machine With Subnetwork Nodes for Representation Learning.

PubMed

Yang, Yimin; Wu, Q M Jonathan

2016-11-01

The extreme learning machine (ELM), which was originally proposed for "generalized" single-hidden layer feedforward neural networks, provides efficient unified learning solutions for the applications of clustering, regression, and classification. It presents competitive accuracy with superb efficiency in many applications. However, ELM with subnetwork nodes architecture has not attracted much research attentions. Recently, many methods have been proposed for supervised/unsupervised dimension reduction or representation learning, but these methods normally only work for one type of problem. This paper studies the general architecture of multilayer ELM (ML-ELM) with subnetwork nodes, showing that: 1) the proposed method provides a representation learning platform with unsupervised/supervised and compressed/sparse representation learning and 2) experimental results on ten image datasets and 16 classification datasets show that, compared to other conventional feature learning methods, the proposed ML-ELM with subnetwork nodes performs competitively or much better than other feature learning methods.
UNMANNED AERIAL VEHICLE (UAV) HYPERSPECTRAL REMOTE SENSING FOR DRYLAND VEGETATION MONITORING

DOE Office of Scientific and Technical Information (OSTI.GOV)

Nancy F. Glenn; Jessica J. Mitchell; Matthew O. Anderson

2012-06-01

UAV-based hyperspectral remote sensing capabilities developed by the Idaho National Lab and Idaho State University, Boise Center Aerospace Lab, were recently tested via demonstration flights that explored the influence of altitude on geometric error, image mosaicking, and dryland vegetation classification. The test flights successfully acquired usable flightline data capable of supporting classifiable composite images. Unsupervised classification results support vegetation management objectives that rely on mapping shrub cover and distribution patterns. Overall, supervised classifications performed poorly despite spectral separability in the image-derived endmember pixels. Future mapping efforts that leverage ground reference data, ultra-high spatial resolution photos and time series analysis shouldmore » be able to effectively distinguish native grasses such as Sandberg bluegrass (Poa secunda), from invasives such as burr buttercup (Ranunculus testiculatus) and cheatgrass (Bromus tectorum).« less
Application of LANDSAT data to monitor land reclamation progress in Belmont County, Ohio

NASA Technical Reports Server (NTRS)

Bloemer, H. H. L.; Brumfield, J. O.; Campbell, W. J.; Witt, R. G.; Bly, B. G.

1981-01-01

Strip and contour mining techniques are reviewed as well as some studies conducted to determine the applicability of LANDSAT and associated digital image processing techniques to the surficial problems associated with mining operations. A nontraditional unsupervised classification approach to multispectral data is considered which renders increased classification separability in land cover analysis of surface mined areas. The approach also reduces the dimensionality of the data and requires only minimal analytical skills in digital data processing.
Wetland delineation with IKONOS high-resolution satellite imagery, Fort Custer Training Center, Battle Creek, Michigan, 2005

USGS Publications Warehouse

Fuller, L.M.; Morgan, T.R.; Aichele, Stephen S.

2006-01-01

The Michigan Army National Guard’s Fort Custer Training Center (FCTC) in Battle Creek, Mich., has the responsibility to protect wetland resources on the training grounds while providing training opportunities, and for future development planning at the facility. The National Wetlands Inventory (NWI) data have been the primary wetland-boundary resource, but a check on scale and accuracy of the wetland boundary information for the Fort Custer Training Center was needed. In cooperation with the FCTC, the U.S. Geological Survey (USGS) used an early spring IKONOS pan-sharpened satellite image to delineate the wetlands and create a more accurate wetland map for the FCTC. The USGS tested automated approaches (supervised and unsupervised classifications) to identify the wetland areas from the IKONOS satellite image, but the automated approaches alone did not yield accurate results. To ensure accurate wetland boundaries, the final wetland map was manually digitized on the basis of the automated supervised and unsupervised classifications, in combination with NWI data, field verifications, and visual interpretation of the IKONOS satellite image. The final wetland areas digitized from the IKONOS satellite imagery were similar to those in NWI; however, the wetland boundaries differed in some areas, a few wetlands mapped on the NWI were determined not to be wetlands from the IKONOS image and field verification, and additional previously unmapped wetlands not recognized by the NWI were identified from the IKONOS image.
Application of diffusion maps to identify human factors of self-reported anomalies in aviation.

PubMed

Andrzejczak, Chris; Karwowski, Waldemar; Mikusinski, Piotr

2012-01-01

A study investigating what factors are present leading to pilots submitting voluntary anomaly reports regarding their flight performance was conducted. Diffusion Maps (DM) were selected as the method of choice for performing dimensionality reduction on text records for this study. Diffusion Maps have seen successful use in other domains such as image classification and pattern recognition. High-dimensionality data in the form of narrative text reports from the NASA Aviation Safety Reporting System (ASRS) were clustered and categorized by way of dimensionality reduction. Supervised analyses were performed to create a baseline document clustering system. Dimensionality reduction techniques identified concepts or keywords within records, and allowed the creation of a framework for an unsupervised document classification system. Results from the unsupervised clustering algorithm performed similarly to the supervised methods outlined in the study. The dimensionality reduction was performed on 100 of the most commonly occurring words within 126,000 text records describing commercial aviation incidents. This study demonstrates that unsupervised machine clustering and organization of incident reports is possible based on unbiased inputs. Findings from this study reinforced traditional views on what factors contribute to civil aviation anomalies, however, new associations between previously unrelated factors and conditions were also found.
Unsupervised classification of remote multispectral sensing data

NASA Technical Reports Server (NTRS)

Su, M. Y.

1972-01-01

The new unsupervised classification technique for classifying multispectral remote sensing data which can be either from the multispectral scanner or digitized color-separation aerial photographs consists of two parts: (a) a sequential statistical clustering which is a one-pass sequential variance analysis and (b) a generalized K-means clustering. In this composite clustering technique, the output of (a) is a set of initial clusters which are input to (b) for further improvement by an iterative scheme. Applications of the technique using an IBM-7094 computer on multispectral data sets over Purdue's Flight Line C-1 and the Yellowstone National Park test site have been accomplished. Comparisons between the classification maps by the unsupervised technique and the supervised maximum liklihood technique indicate that the classification accuracies are in agreement.
Interactive classification and content-based retrieval of tissue images

NASA Astrophysics Data System (ADS)

Aksoy, Selim; Marchisio, Giovanni B.; Tusk, Carsten; Koperski, Krzysztof

2002-11-01

We describe a system for interactive classification and retrieval of microscopic tissue images. Our system models tissues in pixel, region and image levels. Pixel level features are generated using unsupervised clustering of color and texture values. Region level features include shape information and statistics of pixel level feature values. Image level features include statistics and spatial relationships of regions. To reduce the gap between low-level features and high-level expert knowledge, we define the concept of prototype regions. The system learns the prototype regions in an image collection using model-based clustering and density estimation. Different tissue types are modeled using spatial relationships of these regions. Spatial relationships are represented by fuzzy membership functions. The system automatically selects significant relationships from training data and builds models which can also be updated using user relevance feedback. A Bayesian framework is used to classify tissues based on these models. Preliminary experiments show that the spatial relationship models we developed provide a flexible and powerful framework for classification and retrieval of tissue images.
Development and application of operational techniques for the inventory and monitoring of resources and uses for the Texas coastal zone

NASA Technical Reports Server (NTRS)

Harwood, P. (Principal Investigator); Finley, R.; Mcculloch, S.; Marphy, D.; Hupp, B.

1976-01-01

The author has identified the following significant results. Image interpretation mapping techniques were successfully applied to test site 5, an area with a semi-arid climate. The land cover/land use classification required further modification. A new program, HGROUP, added to the ADP classification schedule provides a convenient method for examining the spectral similarity between classes. This capability greatly simplifies the task of combining 25-30 unsupervised subclasses into about 15 major classes that approximately correspond to the land use/land cover classification scheme.
Supervised versus unsupervised categorization: two sides of the same coin?

PubMed

Pothos, Emmanuel M; Edwards, Darren J; Perlman, Amotz

2011-09-01

Supervised and unsupervised categorization have been studied in separate research traditions. A handful of studies have attempted to explore a possible convergence between the two. The present research builds on these studies, by comparing the unsupervised categorization results of Pothos et al. ( 2011 ; Pothos et al., 2008 ) with the results from two procedures of supervised categorization. In two experiments, we tested 375 participants with nine different stimulus sets and examined the relation between ease of learning of a classification, memory for a classification, and spontaneous preference for a classification. After taking into account the role of the number of category labels (clusters) in supervised learning, we found the three variables to be closely associated with each other. Our results provide encouragement for researchers seeking unified theoretical explanations for supervised and unsupervised categorization, but raise a range of challenging theoretical questions.
Unsupervised active learning based on hierarchical graph-theoretic clustering.

PubMed

Hu, Weiming; Hu, Wei; Xie, Nianhua; Maybank, Steve

2009-10-01

Most existing active learning approaches are supervised. Supervised active learning has the following problems: inefficiency in dealing with the semantic gap between the distribution of samples in the feature space and their labels, lack of ability in selecting new samples that belong to new categories that have not yet appeared in the training samples, and lack of adaptability to changes in the semantic interpretation of sample categories. To tackle these problems, we propose an unsupervised active learning framework based on hierarchical graph-theoretic clustering. In the framework, two promising graph-theoretic clustering algorithms, namely, dominant-set clustering and spectral clustering, are combined in a hierarchical fashion. Our framework has some advantages, such as ease of implementation, flexibility in architecture, and adaptability to changes in the labeling. Evaluations on data sets for network intrusion detection, image classification, and video classification have demonstrated that our active learning framework can effectively reduce the workload of manual classification while maintaining a high accuracy of automatic classification. It is shown that, overall, our framework outperforms the support-vector-machine-based supervised active learning, particularly in terms of dealing much more efficiently with new samples whose categories have not yet appeared in the training samples.
Unsupervised Classification of PolSAR Data Using a Scattering Similarity Measure Derived From a Geodesic Distance

NASA Astrophysics Data System (ADS)

Ratha, Debanshu; Bhattacharya, Avik; Frery, Alejandro C.

2018-01-01

In this letter, we propose a novel technique for obtaining scattering components from Polarimetric Synthetic Aperture Radar (PolSAR) data using the geodesic distance on the unit sphere. This geodesic distance is obtained between an elementary target and the observed Kennaugh matrix, and it is further utilized to compute a similarity measure between scattering mechanisms. The normalized similarity measure for each elementary target is then modulated with the total scattering power (Span). This measure is used to categorize pixels into three categories i.e. odd-bounce, double-bounce and volume, depending on which of the above scattering mechanisms dominate. Then the maximum likelihood classifier of [J.-S. Lee, M. R. Grunes, E. Pottier, and L. Ferro-Famil, Unsupervised terrain classification preserving polarimetric scattering characteristics, IEEE Trans. Geos. Rem. Sens., vol. 42, no. 4, pp. 722731, April 2004.] based on the complex Wishart distribution is iteratively used for each category. Dominant scattering mechanisms are thus preserved in this classification scheme. We show results for L-band AIRSAR and ALOS-2 datasets acquired over San Francisco and Mumbai, respectively. The scattering mechanisms are better preserved using the proposed methodology than the unsupervised classification results using the Freeman-Durden scattering powers on an orientation angle (OA) corrected PolSAR image. Furthermore, (1) the scattering similarity is a completely non-negative quantity unlike the negative powers that might occur in double- bounce and odd-bounce scattering component under Freeman Durden decomposition (FDD), and (2) the methodology can be extended to more canonical targets as well as for bistatic scattering.
Biophysical control of intertidal benthic macroalgae revealed by high-frequency multispectral camera images

NASA Astrophysics Data System (ADS)

van der Wal, Daphne; van Dalen, Jeroen; Wielemaker-van den Dool, Annette; Dijkstra, Jasper T.; Ysebaert, Tom

2014-07-01

Intertidal benthic macroalgae are a biological quality indicator in estuaries and coasts. While remote sensing has been applied to quantify the spatial distribution of such macroalgae, it is generally not used for their monitoring. We examined the day-to-day and seasonal dynamics of macroalgal cover on a sandy intertidal flat using visible and near-infrared images from a time-lapse camera mounted on a tower. Benthic algae were identified using supervised, semi-supervised and unsupervised classification techniques, validated with monthly ground-truthing over one year. A supervised classification (based on maximum likelihood, using training areas identified in the field) performed best in discriminating between sediment, benthic diatom films and macroalgae, with highest spectral separability between macroalgae and diatoms in spring/summer. An automated unsupervised classification (based on the Normalised Differential Vegetation Index NDVI) allowed detection of daily changes in macroalgal coverage without the need for calibration. This method showed a bloom of macroalgae (filamentous green algae, Ulva sp.) in summer with > 60% cover, but with pronounced superimposed day-to-day variation in cover. Waves were a major factor in regulating macroalgal cover, but regrowth of the thalli after a summer storm was fast (2 weeks). Images and in situ data demonstrated that the protruding tubes of the polychaete Lanice conchilega facilitated both settlement (anchorage) and survival (resistance to waves) of the macroalgae. Thus, high-frequency, high resolution images revealed the mechanisms for regulating the dynamics in cover of the macroalgae and for their spatial structuring. Ramifications for the mode, timing, frequency and evaluation of monitoring macroalgae by field and remote sensing surveys are discussed.
High and low density development in Puerto Rico

Treesearch

William A. Gould; Sebastian Martinuzzi; Olga M. Ramos Gonzalez

2008-01-01

This map shows the distribution of high and low density developed lands in Puerto Rico (Martinuzzi et al. 2007). The map was created using a mosaic of Landsat ETM+ images that range from the years 2000 to 2003. The developed land cover was classified using the Iterative Self-Organizing Data Analysis Technique (ISODATA) unsupervised classification (ERDAS 2003)....

Assessing Hurricane Katrina Damage to the Mississippi Gulf Coast Using IKONOS Imagery

NASA Technical Reports Server (NTRS)

Spruce, Joseph; McKellip, Rodney

2006-01-01

Hurricane Katrina hit southeastern Louisiana and the Mississippi Gulf Coast as a Category 3 hurricane with storm surges as high as 9 m. Katrina devastated several coastal towns by destroying or severely damaging hundreds of homes. Several Federal agencies are assessing storm impacts and assisting recovery using high-spatial-resolution remotely sensed data from satellite and airborne platforms. High-quality IKONOS satellite imagery was collected on September 2, 2005, over southwestern Mississippi. Pan-sharpened IKONOS multispectral data and ERDAS IMAGINE software were used to classify post-storm land cover for coastal Hancock and Harrison Counties. This classification included a storm debris category of interest to FEMA for disaster mitigation. The classification resulted from combining traditional unsupervised and supervised classification techniques. Higher spatial resolution aerial and handheld photography were used as reference data. Results suggest that traditional classification techniques and IKONOS data can map wood-dominated storm debris in open areas if relevant training areas are used to develop the unsupervised classification signatures. IKONOS data also enabled other hurricane damage assessment, such as flood-deposited mud on lawns and vegetation foliage loss from the storm. IKONOS data has also aided regional Katrina vegetation damage surveys from multidate Land Remote Sensing Satellite and Moderate Resolution Imaging Spectroradiometer data.
Image quality classification for DR screening using deep learning.

PubMed

FengLi Yu; Jing Sun; Annan Li; Jun Cheng; Cheng Wan; Jiang Liu

2017-07-01

The quality of input images significantly affects the outcome of automated diabetic retinopathy (DR) screening systems. Unlike the previous methods that only consider simple low-level features such as hand-crafted geometric and structural features, in this paper we propose a novel method for retinal image quality classification (IQC) that performs computational algorithms imitating the working of the human visual system. The proposed algorithm combines unsupervised features from saliency map and supervised features coming from convolutional neural networks (CNN), which are fed to an SVM to automatically detect high quality vs poor quality retinal fundus images. We demonstrate the superior performance of our proposed algorithm on a large retinal fundus image dataset and the method could achieve higher accuracy than other methods. Although retinal images are used in this study, the methodology is applicable to the image quality assessment and enhancement of other types of medical images.
Semi-supervised manifold learning with affinity regularization for Alzheimer's disease identification using positron emission tomography imaging.

PubMed

Lu, Shen; Xia, Yong; Cai, Tom Weidong; Feng, David Dagan

2015-01-01

Dementia, Alzheimer's disease (AD) in particular is a global problem and big threat to the aging population. An image based computer-aided dementia diagnosis method is needed to providing doctors help during medical image examination. Many machine learning based dementia classification methods using medical imaging have been proposed and most of them achieve accurate results. However, most of these methods make use of supervised learning requiring fully labeled image dataset, which usually is not practical in real clinical environment. Using large amount of unlabeled images can improve the dementia classification performance. In this study we propose a new semi-supervised dementia classification method based on random manifold learning with affinity regularization. Three groups of spatial features are extracted from positron emission tomography (PET) images to construct an unsupervised random forest which is then used to regularize the manifold learning objective function. The proposed method, stat-of-the-art Laplacian support vector machine (LapSVM) and supervised SVM are applied to classify AD and normal controls (NC). The experiment results show that learning with unlabeled images indeed improves the classification performance. And our method outperforms LapSVM on the same dataset.
An Evaluation of Feature Learning Methods for High Resolution Image Classification

NASA Astrophysics Data System (ADS)

Tokarczyk, P.; Montoya, J.; Schindler, K.

2012-07-01

Automatic image classification is one of the fundamental problems of remote sensing research. The classification problem is even more challenging in high-resolution images of urban areas, where the objects are small and heterogeneous. Two questions arise, namely which features to extract from the raw sensor data to capture the local radiometry and image structure at each pixel or segment, and which classification method to apply to the feature vectors. While classifiers are nowadays well understood, selecting the right features remains a largely empirical process. Here we concentrate on the features. Several methods are evaluated which allow one to learn suitable features from unlabelled image data by analysing the image statistics. In a comparative study, we evaluate unsupervised feature learning with different linear and non-linear learning methods, including principal component analysis (PCA) and deep belief networks (DBN). We also compare these automatically learned features with popular choices of ad-hoc features including raw intensity values, standard combinations like the NDVI, a few PCA channels, and texture filters. The comparison is done in a unified framework using the same images, the target classes, reference data and a Random Forest classifier.
Image fusion using sparse overcomplete feature dictionaries

DOEpatents

Brumby, Steven P.; Bettencourt, Luis; Kenyon, Garrett T.; Chartrand, Rick; Wohlberg, Brendt

2015-10-06

Approaches for deciding what individuals in a population of visual system "neurons" are looking for using sparse overcomplete feature dictionaries are provided. A sparse overcomplete feature dictionary may be learned for an image dataset and a local sparse representation of the image dataset may be built using the learned feature dictionary. A local maximum pooling operation may be applied on the local sparse representation to produce a translation-tolerant representation of the image dataset. An object may then be classified and/or clustered within the translation-tolerant representation of the image dataset using a supervised classification algorithm and/or an unsupervised clustering algorithm.
Using deep learning in image hyper spectral segmentation, classification, and detection

NASA Astrophysics Data System (ADS)

Zhao, Xiuying; Su, Zhenyu

2018-02-01

Recent years have shown that deep learning neural networks are a valuable tool in the field of computer vision. Deep learning method can be used in applications like remote sensing such as Land cover Classification, Detection of Vehicle in Satellite Images, Hyper spectral Image classification. This paper addresses the use of the deep learning artificial neural network in Satellite image segmentation. Image segmentation plays an important role in image processing. The hue of the remote sensing image often has a large hue difference, which will result in the poor display of the images in the VR environment. Image segmentation is a pre processing technique applied to the original images and splits the image into many parts which have different hue to unify the color. Several computational models based on supervised, unsupervised, parametric, probabilistic region based image segmentation techniques have been proposed. Recently, one of the machine learning technique known as, deep learning with convolution neural network has been widely used for development of efficient and automatic image segmentation models. In this paper, we focus on study of deep neural convolution network and its variants for automatic image segmentation rather than traditional image segmentation strategies.
Comparison of remote sensing image processing techniques to identify tornado damage areas from Landsat TM data

USGS Publications Warehouse

Myint, S.W.; Yuan, M.; Cerveny, R.S.; Giri, C.P.

2008-01-01

Remote sensing techniques have been shown effective for large-scale damage surveys after a hazardous event in both near real-time or post-event analyses. The paper aims to compare accuracy of common imaging processing techniques to detect tornado damage tracks from Landsat TM data. We employed the direct change detection approach using two sets of images acquired before and after the tornado event to produce a principal component composite images and a set of image difference bands. Techniques in the comparison include supervised classification, unsupervised classification, and objectoriented classification approach with a nearest neighbor classifier. Accuracy assessment is based on Kappa coefficient calculated from error matrices which cross tabulate correctly identified cells on the TM image and commission and omission errors in the result. Overall, the Object-oriented Approach exhibits the highest degree of accuracy in tornado damage detection. PCA and Image Differencing methods show comparable outcomes. While selected PCs can improve detection accuracy 5 to 10%, the Object-oriented Approach performs significantly better with 15-20% higher accuracy than the other two techniques. ?? 2008 by MDPI.
Comparison of Remote Sensing Image Processing Techniques to Identify Tornado Damage Areas from Landsat TM Data

PubMed Central

Myint, Soe W.; Yuan, May; Cerveny, Randall S.; Giri, Chandra P.

2008-01-01

Remote sensing techniques have been shown effective for large-scale damage surveys after a hazardous event in both near real-time or post-event analyses. The paper aims to compare accuracy of common imaging processing techniques to detect tornado damage tracks from Landsat TM data. We employed the direct change detection approach using two sets of images acquired before and after the tornado event to produce a principal component composite images and a set of image difference bands. Techniques in the comparison include supervised classification, unsupervised classification, and object-oriented classification approach with a nearest neighbor classifier. Accuracy assessment is based on Kappa coefficient calculated from error matrices which cross tabulate correctly identified cells on the TM image and commission and omission errors in the result. Overall, the Object-oriented Approach exhibits the highest degree of accuracy in tornado damage detection. PCA and Image Differencing methods show comparable outcomes. While selected PCs can improve detection accuracy 5 to 10%, the Object-oriented Approach performs significantly better with 15-20% higher accuracy than the other two techniques. PMID:27879757
Pulsed terahertz imaging of breast cancer in freshly excised murine tumors

NASA Astrophysics Data System (ADS)

Bowman, Tyler; Chavez, Tanny; Khan, Kamrul; Wu, Jingxian; Chakraborty, Avishek; Rajaram, Narasimhan; Bailey, Keith; El-Shenawee, Magda

2018-02-01

This paper investigates terahertz (THz) imaging and classification of freshly excised murine xenograft breast cancer tumors. These tumors are grown via injection of E0771 breast adenocarcinoma cells into the flank of mice maintained on high-fat diet. Within 1 h of excision, the tumor and adjacent tissues are imaged using a pulsed THz system in the reflection mode. The THz images are classified using a statistical Bayesian mixture model with unsupervised and supervised approaches. Correlation with digitized pathology images is conducted using classification images assigned by a modal class decision rule. The corresponding receiver operating characteristic curves are obtained based on the classification results. A total of 13 tumor samples obtained from 9 tumors are investigated. The results show good correlation of THz images with pathology results in all samples of cancer and fat tissues. For tumor samples of cancer, fat, and muscle tissues, THz images show reasonable correlation with pathology where the primary challenge lies in the overlapping dielectric properties of cancer and muscle tissues. The use of a supervised regression approach shows improvement in the classification images although not consistently in all tissue regions. Advancing THz imaging of breast tumors from mice and the development of accurate statistical models will ultimately progress the technique for the assessment of human breast tumor margins.
Image processing developments and applications for water quality monitoring and trophic state determination

NASA Technical Reports Server (NTRS)

Blackwell, R. J.

1982-01-01

Remote sensing data analysis of water quality monitoring is evaluated. Data anaysis and image processing techniques are applied to LANDSAT remote sensing data to produce an effective operational tool for lake water quality surveying and monitoring. Digital image processing and analysis techniques were designed, developed, tested, and applied to LANDSAT multispectral scanner (MSS) data and conventional surface acquired data. Utilization of these techniques facilitates the surveying and monitoring of large numbers of lakes in an operational manner. Supervised multispectral classification, when used in conjunction with surface acquired water quality indicators, is used to characterize water body trophic status. Unsupervised multispectral classification, when interpreted by lake scientists familiar with a specific water body, yields classifications of equal validity with supervised methods and in a more cost effective manner. Image data base technology is used to great advantage in characterizing other contributing effects to water quality. These effects include drainage basin configuration, terrain slope, soil, precipitation and land cover characteristics.
Using hyperspectral remote sensing for land cover classification

NASA Astrophysics Data System (ADS)

Zhang, Wendy W.; Sriharan, Shobha

2005-01-01

This project used hyperspectral data set to classify land cover using remote sensing techniques. Many different earth-sensing satellites, with diverse sensors mounted on sophisticated platforms, are currently in earth orbit. These sensors are designed to cover a wide range of the electromagnetic spectrum and are generating enormous amounts of data that must be processed, stored, and made available to the user community. The Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) collects data in 224 bands that are approximately 9.6 nm wide in contiguous bands between 0.40 and 2.45 mm. Hyperspectral sensors acquire images in many, very narrow, contiguous spectral bands throughout the visible, near-IR, and thermal IR portions of the spectrum. The unsupervised image classification procedure automatically categorizes the pixels in an image into land cover classes or themes. Experiments on using hyperspectral remote sensing for land cover classification were conducted during the 2003 and 2004 NASA Summer Faculty Fellowship Program at Stennis Space Center. Research Systems Inc.'s (RSI) ENVI software package was used in this application framework. In this application, emphasis was placed on: (1) Spectrally oriented classification procedures for land cover mapping, particularly, the supervised surface classification using AVIRIS data; and (2) Identifying data endmembers.
Combining Unsupervised and Supervised Classification to Build User Models for Exploratory Learning Environments

ERIC Educational Resources Information Center

Amershi, Saleema; Conati, Cristina

2009-01-01

In this paper, we present a data-based user modeling framework that uses both unsupervised and supervised classification to build student models for exploratory learning environments. We apply the framework to build student models for two different learning environments and using two different data sources (logged interface and eye-tracking data).…
Early breast tumor and late SARS detections using space-variant multispectral infrared imaging at a single pixel

NASA Astrophysics Data System (ADS)

Szu, Harold H.; Buss, James R.; Kopriva, Ivica

2004-04-01

We proposed the physics approach to solve a physical inverse problem, namely to choose the unique equilibrium solution (at the minimum free energy: H= E - ToS, including the Wiener, l.m.s E, and ICA, Max S, as special cases). The "unsupervised classification" presumes that required information must be learned and derived directly and solely from the data alone, in consistence with the classical Duda-Hart ATR definition of the "unlabelled data". Such truly unsupervised methodology is presented for space-variant imaging processing for a single pixel in the real world case of remote sensing, early tumor detections and SARS. The indeterminacy of the multiple solutions of the inverse problem is regulated or selected by means of the absolute minimum of isothermal free energy as the ground truth of local equilibrium condition at the single-pixel foot print.
Applications of LANDSAT data to the integrated economic development of Mindoro, Phillipines

NASA Technical Reports Server (NTRS)

Wagner, T. W.; Fernandez, J. C.

1977-01-01

LANDSAT data is seen as providing essential up-to-date resource information for the planning process. LANDSAT data of Mindoro Island in the Philippines was processed to provide thematic maps showing patterns of agriculture, forest cover, terrain, wetlands and water turbidity. A hybrid approach using both supervised and unsupervised classification techniques resulted in 30 different scene classes which were subsequently color-coded and mapped at a scale of 1:250,000. In addition, intensive image analysis is being carried out in evaluating the images. The images, maps, and aerial statistics are being used to provide data to seven technical departments in planning the economic development of Mindoro. Multispectral aircraft imagery was collected to compliment the application of LANDSAT data and validate the classification results.
Do deep convolutional neural networks really need to be deep when applied for remote scene classification?

NASA Astrophysics Data System (ADS)

Luo, Chang; Wang, Jie; Feng, Gang; Xu, Suhui; Wang, Shiqiang

2017-10-01

Deep convolutional neural networks (CNNs) have been widely used to obtain high-level representation in various computer vision tasks. However, for remote scene classification, there are not sufficient images to train a very deep CNN from scratch. From two viewpoints of generalization power, we propose two promising kinds of deep CNNs for remote scenes and try to find whether deep CNNs need to be deep for remote scene classification. First, we transfer successful pretrained deep CNNs to remote scenes based on the theory that depth of CNNs brings the generalization power by learning available hypothesis for finite data samples. Second, according to the opposite viewpoint that generalization power of deep CNNs comes from massive memorization and shallow CNNs with enough neural nodes have perfect finite sample expressivity, we design a lightweight deep CNN (LDCNN) for remote scene classification. With five well-known pretrained deep CNNs, experimental results on two independent remote-sensing datasets demonstrate that transferred deep CNNs can achieve state-of-the-art results in an unsupervised setting. However, because of its shallow architecture, LDCNN cannot obtain satisfactory performance, regardless of whether in an unsupervised, semisupervised, or supervised setting. CNNs really need depth to obtain general features for remote scenes. This paper also provides baseline for applying deep CNNs to other remote sensing tasks.
Unifying framework for multimodal brain MRI segmentation based on Hidden Markov Chains.

PubMed

Bricq, S; Collet, Ch; Armspach, J P

2008-12-01

In the frame of 3D medical imaging, accurate segmentation of multimodal brain MR images is of interest for many brain disorders. However, due to several factors such as noise, imaging artifacts, intrinsic tissue variation and partial volume effects, tissue classification remains a challenging task. In this paper, we present a unifying framework for unsupervised segmentation of multimodal brain MR images including partial volume effect, bias field correction, and information given by a probabilistic atlas. Here-proposed method takes into account neighborhood information using a Hidden Markov Chain (HMC) model. Due to the limited resolution of imaging devices, voxels may be composed of a mixture of different tissue types, this partial volume effect is included to achieve an accurate segmentation of brain tissues. Instead of assigning each voxel to a single tissue class (i.e., hard classification), we compute the relative amount of each pure tissue class in each voxel (mixture estimation). Further, a bias field estimation step is added to the proposed algorithm to correct intensity inhomogeneities. Furthermore, atlas priors were incorporated using probabilistic brain atlas containing prior expectations about the spatial localization of different tissue classes. This atlas is considered as a complementary sensor and the proposed method is extended to multimodal brain MRI without any user-tunable parameter (unsupervised algorithm). To validate this new unifying framework, we present experimental results on both synthetic and real brain images, for which the ground truth is available. Comparison with other often used techniques demonstrates the accuracy and the robustness of this new Markovian segmentation scheme.
Modeling Image Patches with a Generic Dictionary of Mini-Epitomes

PubMed Central

Papandreou, George; Chen, Liang-Chieh; Yuille, Alan L.

2015-01-01

The goal of this paper is to question the necessity of features like SIFT in categorical visual recognition tasks. As an alternative, we develop a generative model for the raw intensity of image patches and show that it can support image classification performance on par with optimized SIFT-based techniques in a bag-of-visual-words setting. Key ingredient of the proposed model is a compact dictionary of mini-epitomes, learned in an unsupervised fashion on a large collection of images. The use of epitomes allows us to explicitly account for photometric and position variability in image appearance. We show that this flexibility considerably increases the capacity of the dictionary to accurately approximate the appearance of image patches and support recognition tasks. For image classification, we develop histogram-based image encoding methods tailored to the epitomic representation, as well as an “epitomic footprint” encoding which is easy to visualize and highlights the generative nature of our model. We discuss in detail computational aspects and develop efficient algorithms to make the model scalable to large tasks. The proposed techniques are evaluated with experiments on the challenging PASCAL VOC 2007 image classification benchmark. PMID:26321859
Multispectral and Panchromatic used Enhancement Resolution and Study Effective Enhancement on Supervised and Unsupervised Classification Land – Cover

NASA Astrophysics Data System (ADS)

Salman, S. S.; Abbas, W. A.

2018-05-01

The goal of the study is to support analysis Enhancement of Resolution and study effect on classification methods on bands spectral information of specific and quantitative approaches. In this study introduce a method to enhancement resolution Landsat 8 of combining the bands spectral of 30 meters resolution with panchromatic band 8 of 15 meters resolution, because of importance multispectral imagery to extracting land - cover. Classification methods used in this study to classify several lands -covers recorded from OLI- 8 imagery. Two methods of Data mining can be classified as either supervised or unsupervised. In supervised methods, there is a particular predefined target, that means the algorithm learn which values of the target are associated with which values of the predictor sample. K-nearest neighbors and maximum likelihood algorithms examine in this work as supervised methods. In other hand, no sample identified as target in unsupervised methods, the algorithm of data extraction searches for structure and patterns between all the variables, represented by Fuzzy C-mean clustering method as one of the unsupervised methods, NDVI vegetation index used to compare the results of classification method, the percent of dense vegetation in maximum likelihood method give a best results.
Extraction of features from medical images using a modular neural network approach that relies on learning by sample

NASA Astrophysics Data System (ADS)

Brahmi, Djamel; Serruys, Camille; Cassoux, Nathalie; Giron, Alain; Triller, Raoul; Lehoang, Phuc; Fertil, Bernard

2000-06-01

Medical images provide experienced physicians with meaningful visual stimuli but their features are frequently hard to decipher. The development of a computational model to mimic physicians' expertise is a demanding task, especially if a significant and sophisticated preprocessing of images is required. Learning from well-expertised images may be a more convenient approach, inasmuch a large and representative bunch of samples is available. A four-stage approach has been designed, which combines image sub-sampling with unsupervised image coding, supervised classification and image reconstruction in order to directly extract medical expertise from raw images. The system has been applied (1) to the detection of some features related to the diagnosis of black tumors of skin (a classification issue) and (2) to the detection of virus-infected and healthy areas in retina angiography in order to locate precisely the border between them and characterize the evolution of infection. For reasonably balanced training sets, we are able to obtained about 90% correct classification of features (black tumors). Boundaries generated by our system mimic reproducibility of hand-outlines drawn by experts (segmentation of virus-infected area).
Automatic Cell Segmentation Using a Shape-Classification Model in Immunohistochemically Stained Cytological Images

NASA Astrophysics Data System (ADS)

Shah, Shishir

This paper presents a segmentation method for detecting cells in immunohistochemically stained cytological images. A two-phase approach to segmentation is used where an unsupervised clustering approach coupled with cluster merging based on a fitness function is used as the first phase to obtain a first approximation of the cell locations. A joint segmentation-classification approach incorporating ellipse as a shape model is used as the second phase to detect the final cell contour. The segmentation model estimates a multivariate density function of low-level image features from training samples and uses it as a measure of how likely each image pixel is to be a cell. This estimate is constrained by the zero level set, which is obtained as a solution to an implicit representation of an ellipse. Results of segmentation are presented and compared to ground truth measurements.

An Automatic Segmentation and Classification Framework Based on PCNN Model for Single Tooth in MicroCT Images.

PubMed

Wang, Liansheng; Li, Shusheng; Chen, Rongzhen; Liu, Sze-Yu; Chen, Jyh-Cheng

2016-01-01

Accurate segmentation and classification of different anatomical structures of teeth from medical images plays an essential role in many clinical applications. Usually, the anatomical structures of teeth are manually labelled by experienced clinical doctors, which is time consuming. However, automatic segmentation and classification is a challenging task because the anatomical structures and surroundings of the tooth in medical images are rather complex. Therefore, in this paper, we propose an effective framework which is designed to segment the tooth with a Selective Binary and Gaussian Filtering Regularized Level Set (GFRLS) method improved by fully utilizing three dimensional (3D) information, and classify the tooth by employing unsupervised learning Pulse Coupled Neural Networks (PCNN) model. In order to evaluate the proposed method, the experiments are conducted on the different datasets of mandibular molars and the experimental results show that our method can achieve better accuracy and robustness compared to other four state of the art clustering methods.
A segmentation and classification scheme for single tooth in MicroCT images based on 3D level set and k-means+.

PubMed

Wang, Liansheng; Li, Shusheng; Chen, Rongzhen; Liu, Sze-Yu; Chen, Jyh-Cheng

2017-04-01

Accurate classification of different anatomical structures of teeth from medical images provides crucial information for the stress analysis in dentistry. Usually, the anatomical structures of teeth are manually labeled by experienced clinical doctors, which is time consuming. However, automatic segmentation and classification is a challenging task because the anatomical structures and surroundings of the tooth in medical images are rather complex. Therefore, in this paper, we propose an effective framework which is designed to segment the tooth with a Selective Binary and Gaussian Filtering Regularized Level Set (GFRLS) method improved by fully utilizing 3 dimensional (3D) information, and classify the tooth by employing unsupervised learning i.e., k-means++ method. In order to evaluate the proposed method, the experiments are conducted on the sufficient and extensive datasets of mandibular molars. The experimental results show that our method can achieve higher accuracy and robustness compared to other three clustering methods. Copyright © 2016 Elsevier Ltd. All rights reserved.
Urban Shanty Town Recognition Based on High-Resolution Remote Sensing Images and National Geographical Monitoring Features - a Case Study of Nanning City

NASA Astrophysics Data System (ADS)

He, Y.; He, Y.

2018-04-01

Urban shanty towns are communities that has contiguous old and dilapidated houses with more than 2000 square meters built-up area or more than 50 households. This study makes attempts to extract shanty towns in Nanning City using the product of Census and TripleSat satellite images. With 0.8-meter high-resolution remote sensing images, five texture characteristics (energy, contrast, maximum probability, and inverse difference moment) of shanty towns are trained and analyzed through GLCM. In this study, samples of shanty town are well classified with 98.2 % producer accuracy of unsupervised classification and 73.2 % supervised classification correctness. Low-rise and mid-rise residential blocks in Nanning City are classified into 4 different types by using k-means clustering and nearest neighbour classification respectively. This study initially establish texture feature descriptions of different types of residential areas, especially low-rise and mid-rise buildings, which would help city administrator evaluate residential blocks and reconstruction shanty towns.
Automated unsupervised multi-parametric classification of adipose tissue depots in skeletal muscle

PubMed Central

Valentinitsch, Alexander; Karampinos, Dimitrios C.; Alizai, Hamza; Subburaj, Karupppasamy; Kumar, Deepak; Link, Thomas M.; Majumdar, Sharmila

2012-01-01

Purpose To introduce and validate an automated unsupervised multi-parametric method for segmentation of the subcutaneous fat and muscle regions in order to determine subcutaneous adipose tissue (SAT) and intermuscular adipose tissue (IMAT) areas based on data from a quantitative chemical shift-based water-fat separation approach. Materials and Methods Unsupervised standard k-means clustering was employed to define sets of similar features (k = 2) within the whole multi-modal image after the water-fat separation. The automated image processing chain was composed of three primary stages including tissue, muscle and bone region segmentation. The algorithm was applied on calf and thigh datasets to compute SAT and IMAT areas and was compared to a manual segmentation. Results The IMAT area using the automatic segmentation had excellent agreement with the IMAT area using the manual segmentation for all the cases in the thigh (R2: 0.96) and for cases with up to moderate IMAT area in the calf (R2: 0.92). The group with the highest grade of muscle fat infiltration in the calf had the highest error in the inner SAT contour calculation. Conclusion The proposed multi-parametric segmentation approach combined with quantitative water-fat imaging provides an accurate and reliable method for an automated calculation of the SAT and IMAT areas reducing considerably the total post-processing time. PMID:23097409
Mapping of Geographically Isolated Wetlands of Western Siberia Using High Resolution Space Images

NASA Astrophysics Data System (ADS)

Dyukarev, E.; Pologova, N.; Dyukarev, A.; Lane, C.; Autrey, B. C.

2014-12-01

Using the remote sensing data for integrated study of natural objects is actual for investigation of difficult to access areas of West Siberia. The research of this study focuses on determining the extent and spectral signatures of isolated wetlands within Ob-Tom Interfluve area using Landsat and Quickbird space images. High-resolution space images were carefully examined and wetlands were manually delineated. Wetlands have clear visible signs at the high resolution space images. 567 wetlands were recognized as isolated wetlands with the area about 10 000 ha (of 2.5% of the study area). Isolated wetlands with area less 2 ha are the most frequent. Half of the total amount of wetlands has area less than 6.4 ha. The largest isolated wetland occupies 797 ha, and only 5% have area more than 50 ha. The Landsat 7 ETM+ data were used for analysis of vegetation structure and spectral characteristics of wetlands. The masked isolated wetlands image was classified into 12 land cover classes using ISODATA unsupervised classification. The attribution of unsupervised classification results allowed us to clearly recognize 7 types of wetlands: tall, low and sparse ryams (Pine-Shrub-Sphagnum community), open wetlands with shrub, moss or sedge cover, and open water objects. Analysis of spectral profiles for all classes has shown that Landsat spectral bands 4 and 5 have higher variability. These bands allow to separate wetland classed definitely. Accuracy assessment of isolated wetland map shows a good agreement with expert field data. The work was supported by grants ISTC № 4079.
Classification of multispectral or hyperspectral satellite imagery using clustering of sparse approximations on sparse representations in learned dictionaries obtained using efficient convolutional sparse coding

DOEpatents

Moody, Daniela; Wohlberg, Brendt

2018-01-02

An approach for land cover classification, seasonal and yearly change detection and monitoring, and identification of changes in man-made features may use a clustering of sparse approximations (CoSA) on sparse representations in learned dictionaries. The learned dictionaries may be derived using efficient convolutional sparse coding to build multispectral or hyperspectral, multiresolution dictionaries that are adapted to regional satellite image data. Sparse image representations of images over the learned dictionaries may be used to perform unsupervised k-means clustering into land cover categories. The clustering process behaves as a classifier in detecting real variability. This approach may combine spectral and spatial textural characteristics to detect geologic, vegetative, hydrologic, and man-made features, as well as changes in these features over time.
Unsupervised classification of variable stars

NASA Astrophysics Data System (ADS)

Valenzuela, Lucas; Pichara, Karim

2018-03-01

During the past 10 years, a considerable amount of effort has been made to develop algorithms for automatic classification of variable stars. That has been primarily achieved by applying machine learning methods to photometric data sets where objects are represented as light curves. Classifiers require training sets to learn the underlying patterns that allow the separation among classes. Unfortunately, building training sets is an expensive process that demands a lot of human efforts. Every time data come from new surveys; the only available training instances are the ones that have a cross-match with previously labelled objects, consequently generating insufficient training sets compared with the large amounts of unlabelled sources. In this work, we present an algorithm that performs unsupervised classification of variable stars, relying only on the similarity among light curves. We tackle the unsupervised classification problem by proposing an untraditional approach. Instead of trying to match classes of stars with clusters found by a clustering algorithm, we propose a query-based method where astronomers can find groups of variable stars ranked by similarity. We also develop a fast similarity function specific for light curves, based on a novel data structure that allows scaling the search over the entire data set of unlabelled objects. Experiments show that our unsupervised model achieves high accuracy in the classification of different types of variable stars and that the proposed algorithm scales up to massive amounts of light curves.
Change detection and classification in brain MR images using change vector analysis.

PubMed

Simões, Rita; Slump, Cornelis

2011-01-01

The automatic detection of longitudinal changes in brain images is valuable in the assessment of disease evolution and treatment efficacy. Most existing change detection methods that are currently used in clinical research to monitor patients suffering from neurodegenerative diseases--such as Alzheimer's--focus on large-scale brain deformations. However, such patients often have other brain impairments, such as infarcts, white matter lesions and hemorrhages, which are typically overlooked by the deformation-based methods. Other unsupervised change detection algorithms have been proposed to detect tissue intensity changes. The outcome of these methods is typically a binary change map, which identifies changed brain regions. However, understanding what types of changes these regions underwent is likely to provide equally important information about lesion evolution. In this paper, we present an unsupervised 3D change detection method based on Change Vector Analysis. We compute and automatically threshold the Generalized Likelihood Ratio map to obtain a binary change map. Subsequently, we perform histogram-based clustering to classify the change vectors. We obtain a Kappa Index of 0.82 using various types of simulated lesions. The classification error is 2%. Finally, we are able to detect and discriminate both small changes and ventricle expansions in datasets from Mild Cognitive Impairment patients.
The dynamics of human-induced land cover change in miombo ecosystems of southern Africa

NASA Astrophysics Data System (ADS)

Jaiteh, Malanding Sambou

Understanding human-induced land cover change in the miombo require the consistent, geographically-referenced, data on temporal land cover characteristics as well as biophysical and socioeconomic drivers of land use, the major cause of land cover change. The overall goal of this research to examine the applications of high-resolution satellite remote sensing data in studying the dynamics of human-induced land cover change in the miombo. Specific objectives are to: (1) evaluate the applications of computer-assisted classification of Landsat Thematic Mapper (TM) data for land cover mapping in the miombo and (2) analyze spatial and temporal patterns of landscape change locations in the miombo. Stepwise Thematic Classification, STC (a hybrid supervised-unsupervised classification) procedure for classifying Landsat TM data was developed and tested using Landsat TM data. Classification accuracy results were compared to those from supervised and unsupervised classification. The STC provided the highest classification accuracy i.e., 83.9% correspondence between classified and referenced data compared to 44.2% and 34.5% for unsupervised and supervised classification respectively. Improvements in the classification process can be attributed to thematic stratification of the image data into spectrally homogenous (thematic) groups and step-by-step classification of the groups using supervised or unsupervised classification techniques. Supervised classification failed to classify 18% of the scene evidence that training data used did not adequately represent all of the variability in the data. Application of the procedure in drier miombo produced overall classification accuracy of 63%. This is much lower than that of wetter miombo. The results clearly demonstrate that digital classification of Landsat TM can be successfully implemented in the miombo without intensive fieldwork. Spatial characteristics of land cover change in agricultural and forested landscapes in central Malawi were analyzed for the period 1984 to 1995 spatial pattern analysis methods. Shifting cultivation areas, Agriculture in forested landscape, experienced highest rate of woodland cover fragmentation with mean patch size of closed woodland cover decreasing from 20ha to 7.5ha. Permanent bare (cropland and settlement) in intensive agricultural matrix landscapes increased 52% largely through the conversion of fallow areas. Protected National Park area remained fairly unchanged although closed woodland area increased by 4%, mainly from regeneration of open woodland. This study provided evidence that changes in spatial characteristics in the miombo differ with landscape. Land use change (i.e. conversion to cropland) is the primary driving force behind changes in landscape spatial patterns. Also, results revealed that exclusion of intense human use (i.e. cultivation and woodcutting) through regulations and/or fencing increased both closed woodland area (through regeneration of open woodland) and overall connectivity in the landscape. Spatial characteristics of land cover change were analyzed at locations in Malawi (wetter miombo) and Zimbabwe (drier miombo). Results indicate land cover dynamics differ both between and within case study sites. In communal areas in the Kasungu scene, land cover change is dominated by woodland fragmentation to open vegetation. Change in private commercial lands was dominantly expansion of bare (settlement and cropland) areas primarily at the expense of open vegetation (fallow land).
Surface mapping via unsupervised classification of remote sensing: application to MESSENGER/MASCS and DAWN/VIRS data.

NASA Astrophysics Data System (ADS)

D'Amore, M.; Le Scaon, R.; Helbert, J.; Maturilli, A.

2017-12-01

Machine-learning achieved unprecedented results in high-dimensional data processing tasks with wide applications in various fields. Due to the growing number of complex nonlinear systems that have to be investigated in science and the bare raw size of data nowadays available, ML offers the unique ability to extract knowledge, regardless the specific application field. Examples are image segmentation, supervised/unsupervised/ semi-supervised classification, feature extraction, data dimensionality analysis/reduction.The MASCS instrument has mapped Mercury surface in the 400-1145 nm wavelength range during orbital observations by the MESSENGER spacecraft. We have conducted k-means unsupervised hierarchical clustering to identify and characterize spectral units from MASCS observations. The results display a dichotomy: a polar and equatorial units, possibly linked to compositional differences or weathering due to irradiation. To explore possible relations between composition and spectral behavior, we have compared the spectral provinces with elemental abundance maps derived from MESSENGER's X-Ray Spectrometer (XRS).For the Vesta application on DAWN Visible and infrared spectrometer (VIR) data, we explored several Machine Learning techniques: image segmentation method, stream algorithm and hierarchical clustering.The algorithm successfully separates the Olivine outcrops around two craters on Vesta's surface [1]. New maps summarizing the spectral and chemical signature of the surface could be automatically produced.We conclude that instead of hand digging in data, scientist could choose a subset of algorithms with well known feature (i.e. efficacy on the particular problem, speed, accuracy) and focus their effort in understanding what important characteristic of the groups found in the data mean. [1] E Ammannito et al. "Olivine in an unexpected location on Vesta's surface". In: Nature 504.7478 (2013), pp. 122-125.
Testing random forest classification for identifying lava flows and mapping age groups on a single Landsat 8 image

NASA Astrophysics Data System (ADS)

Li, Long; Solana, Carmen; Canters, Frank; Kervyn, Matthieu

2017-10-01

Mapping lava flows using satellite images is an important application of remote sensing in volcanology. Several volcanoes have been mapped through remote sensing using a wide range of data, from optical to thermal infrared and radar images, using techniques such as manual mapping, supervised/unsupervised classification, and elevation subtraction. So far, spectral-based mapping applications mainly focus on the use of traditional pixel-based classifiers, without much investigation into the added value of object-based approaches and into advantages of using machine learning algorithms. In this study, Nyamuragira, characterized by a series of > 20 overlapping lava flows erupted over the last century, was used as a case study. The random forest classifier was tested to map lava flows based on pixels and objects. Image classification was conducted for the 20 individual flows and for 8 groups of flows of similar age using a Landsat 8 image and a DEM of the volcano, both at 30-meter spatial resolution. Results show that object-based classification produces maps with continuous and homogeneous lava surfaces, in agreement with the physical characteristics of lava flows, while lava flows mapped through the pixel-based classification are heterogeneous and fragmented including much "salt and pepper noise". In terms of accuracy, both pixel-based and object-based classification performs well but the former results in higher accuracies than the latter except for mapping lava flow age groups without using topographic features. It is concluded that despite spectral similarity, lava flows of contrasting age can be well discriminated and mapped by means of image classification. The classification approach demonstrated in this study only requires easily accessible image data and can be applied to other volcanoes as well if there is sufficient information to calibrate the mapping.
Class imbalance in unsupervised change detection - A diagnostic analysis from urban remote sensing

NASA Astrophysics Data System (ADS)

Leichtle, Tobias; Geiß, Christian; Lakes, Tobia; Taubenböck, Hannes

2017-08-01

Automatic monitoring of changes on the Earth's surface is an intrinsic capability and simultaneously a persistent methodological challenge in remote sensing, especially regarding imagery with very-high spatial resolution (VHR) and complex urban environments. In order to enable a high level of automatization, the change detection problem is solved in an unsupervised way to alleviate efforts associated with collection of properly encoded prior knowledge. In this context, this paper systematically investigates the nature and effects of class distribution and class imbalance in an unsupervised binary change detection application based on VHR imagery over urban areas. For this purpose, a diagnostic framework for sensitivity analysis of a large range of possible degrees of class imbalance is presented, which is of particular importance with respect to unsupervised approaches where the content of images and thus the occurrence and the distribution of classes are generally unknown a priori. Furthermore, this framework can serve as a general technique to evaluate model transferability in any two-class classification problem. The applied change detection approach is based on object-based difference features calculated from VHR imagery and subsequent unsupervised two-class clustering using k-means, genetic k-means and self-organizing map (SOM) clustering. The results from two test sites with different structural characteristics of the built environment demonstrated that classification performance is generally worse in imbalanced class distribution settings while best results were reached in balanced or close to balanced situations. Regarding suitable accuracy measures for evaluating model performance in imbalanced settings, this study revealed that the Kappa statistics show significant response to class distribution while the true skill statistic was widely insensitive to imbalanced classes. In general, the genetic k-means clustering algorithm achieved the most robust results with respect to class imbalance while the SOM clustering exhibited a distinct optimization towards a balanced distribution of classes.
Analysis of the changes in the tarcrete layer on the desert surface of Kuwait using satellite imagery and cell-based modeling

NASA Astrophysics Data System (ADS)

Al-Doasari, Ahmad E.

The 1991 Gulf War caused massive environmental damage in Kuwait. Deposition of oil and soot droplets from hundreds of burning oil-wells created a layer of tarcrete on the desert surface covering over 900 km2. This research investigates the spatial change in the tarcrete extent from 1991 to 1998 using Landsat Thematic Mapper (TM) imagery and statistical modeling techniques. The pixel structure of TM data allows the spatial analysis of the change in tarcrete extent to be conducted at the pixel (cell) level within a geographical information system (GIS). There are two components to this research. The first is a comparison of three remote sensing classification techniques used to map the tarcrete layer. The second is a spatial-temporal analysis and simulation of tarcrete changes through time. The analysis focuses on an area of 389 km2 located south of the Al-Burgan oil field. Five TM images acquired in 1991, 1993, 1994, 1995, and 1998 were geometrically and atmospherically corrected. These images were classified into six classes: oil lakes; heavy, intermediate, light, and traces of tarcrete; and sand. The classification methods tested were unsupervised, supervised, and neural network supervised (fuzzy ARTMAP). Field data of tarcrete characteristics were collected to support the classification process and to evaluate the classification accuracies. Overall, the neural network method is more accurate (60 percent) than the other two methods; both the unsupervised and the supervised classification accuracy assessments resulted in 46 percent accuracy. The five classifications were used in a lagged autologistic model to analyze the spatial changes of the tarcrete through time. The autologistic model correctly identified overall tarcrete contraction between 1991--1993 and 1995--1998. However, tarcrete contraction between 1993--1994 and 1994--1995 was less well marked, in part because of classification errors in the maps from these time periods. Initial simulations of tarcrete contraction with a cellular automaton model were not very successful. However, more accurate classifications could improve the simulations. This study illustrates how an empirical investigation using satellite images, field data, GIS, and spatial statistics can simulate dynamic land-cover change through the use of a discrete statistical and cellular automaton model.
LANDSAT landcover information applied to regional planning decisions. [Prince Edward County, Virginia

NASA Technical Reports Server (NTRS)

Dixon, C. M.

1981-01-01

Land cover information derived from LANDSAT is being utilized by Piedmont Planning District Commission located in the State of Virginia. Progress to date is reported on a level one land cover classification map being produced with nine categories. The nine categories of classification are defined. The computer compatible tape selection is presented. Two unsupervised classifications were done, with 50 and 70 classes respectively. Twenty-eight spectral classes were developed using the supervised technique, employing actual ground truth training sites. The accuracy of the unsupervised classifications are estimated through comparison with local county statistics and with an actual pixel count of LANDSAT information compared to ground truth.
Evaluating pixel and object based image classification techniques for mapping plant invasions from UAV derived aerial imagery: Harrisia pomanensis as a case study

NASA Astrophysics Data System (ADS)

Mafanya, Madodomzi; Tsele, Philemon; Botai, Joel; Manyama, Phetole; Swart, Barend; Monate, Thabang

2017-07-01

Invasive alien plants (IAPs) not only pose a serious threat to biodiversity and water resources but also have impacts on human and animal wellbeing. To support decision making in IAPs monitoring, semi-automated image classifiers which are capable of extracting valuable information in remotely sensed data are vital. This study evaluated the mapping accuracies of supervised and unsupervised image classifiers for mapping Harrisia pomanensis (a cactus plant commonly known as the Midnight Lady) using two interlinked evaluation strategies i.e. point and area based accuracy assessment. Results of the point-based accuracy assessment show that with reference to 219 ground control points, the supervised image classifiers (i.e. Maxver and Bhattacharya) mapped H. pomanensis better than the unsupervised image classifiers (i.e. K-mediuns, Euclidian Length and Isoseg). In this regard, user and producer accuracies were 82.4% and 84% respectively for the Maxver classifier. The user and producer accuracies for the Bhattacharya classifier were 90% and 95.7%, respectively. Though the Maxver produced a higher overall accuracy and Kappa estimate than the Bhattacharya classifier, the Maxver Kappa estimate of 0.8305 is not significantly (statistically) greater than the Bhattacharya Kappa estimate of 0.8088 at a 95% confidence interval. The area based accuracy assessment results show that the Bhattacharya classifier estimated the spatial extent of H. pomanensis with an average mapping accuracy of 86.1% whereas the Maxver classifier only gave an average mapping accuracy of 65.2%. Based on these results, the Bhattacharya classifier is therefore recommended for mapping H. pomanensis. These findings will aid in the algorithm choice making for the development of a semi-automated image classification system for mapping IAPs.
Development Of Polarimetric Decomposition Techniques For Indian Forest Resource Assessment Using Radar Imaging Satellite (Risat-1) Images

NASA Astrophysics Data System (ADS)

Sridhar, J.

2015-12-01

The focus of this work is to examine polarimetric decomposition techniques primarily focussed on Pauli decomposition and Sphere Di-Plane Helix (SDH) decomposition for forest resource assessment. The data processing methods adopted are Pre-processing (Geometric correction and Radiometric calibration), Speckle Reduction, Image Decomposition and Image Classification. Initially to classify forest regions, unsupervised classification was applied to determine different unknown classes. It was observed K-means clustering method gave better results in comparison with ISO Data method.Using the algorithm developed for Radar Tools, the code for decomposition and classification techniques were applied in Interactive Data Language (IDL) and was applied to RISAT-1 image of Mysore-Mandya region of Karnataka, India. This region is chosen for studying forest vegetation and consists of agricultural lands, water and hilly regions. Polarimetric SAR data possess a high potential for classification of earth surface.After applying the decomposition techniques, classification was done by selecting region of interests andpost-classification the over-all accuracy was observed to be higher in the SDH decomposed image, as it operates on individual pixels on a coherent basis and utilises the complete intrinsic coherent nature of polarimetric SAR data. Thereby, making SDH decomposition particularly suited for analysis of high-resolution SAR data. The Pauli Decomposition represents all the polarimetric information in a single SAR image however interpretation of the resulting image is difficult. The SDH decomposition technique seems to produce better results and interpretation as compared to Pauli Decomposition however more quantification and further analysis are being done in this area of research. The comparison of Polarimetric decomposition techniques and evolutionary classification techniques will be the scope of this work.
Unsupervised segmentation of lungs from chest radiographs

NASA Astrophysics Data System (ADS)

Ghosh, Payel; Antani, Sameer K.; Long, L. Rodney; Thoma, George R.

2012-03-01

This paper describes our preliminary investigations for deriving and characterizing coarse-level textural regions present in the lung field on chest radiographs using unsupervised grow-cut (UGC), a cellular automaton based unsupervised segmentation technique. The segmentation has been performed on a publicly available data set of chest radiographs. The algorithm is useful for this application because it automatically converges to a natural segmentation of the image from random seed points using low-level image features such as pixel intensity values and texture features. Our goal is to develop a portable screening system for early detection of lung diseases for use in remote areas in developing countries. This involves developing automated algorithms for screening x-rays as normal/abnormal with a high degree of sensitivity, and identifying lung disease patterns on chest x-rays. Automatically deriving and quantitatively characterizing abnormal regions present in the lung field is the first step toward this goal. Therefore, region-based features such as geometrical and pixel-value measurements were derived from the segmented lung fields. In the future, feature selection and classification will be performed to identify pathological conditions such as pulmonary tuberculosis on chest radiographs. Shape-based features will also be incorporated to account for occlusions of the lung field and by other anatomical structures such as the heart and diaphragm.
Misclassification Errors in Unsupervised Classification Methods. Comparison Based on the Simulation of Targeted Proteomics Data

PubMed Central

Andreev, Victor P; Gillespie, Brenda W; Helfand, Brian T; Merion, Robert M

2016-01-01

Unsupervised classification methods are gaining acceptance in omics studies of complex common diseases, which are often vaguely defined and are likely the collections of disease subtypes. Unsupervised classification based on the molecular signatures identified in omics studies have the potential to reflect molecular mechanisms of the subtypes of the disease and to lead to more targeted and successful interventions for the identified subtypes. Multiple classification algorithms exist but none is ideal for all types of data. Importantly, there are no established methods to estimate sample size in unsupervised classification (unlike power analysis in hypothesis testing). Therefore, we developed a simulation approach allowing comparison of misclassification errors and estimating the required sample size for a given effect size, number, and correlation matrix of the differentially abundant proteins in targeted proteomics studies. All the experiments were performed in silico. The simulated data imitated the expected one from the study of the plasma of patients with lower urinary tract dysfunction with the aptamer proteomics assay Somascan (SomaLogic Inc, Boulder, CO), which targeted 1129 proteins, including 330 involved in inflammation, 180 in stress response, 80 in aging, etc. Three popular clustering methods (hierarchical, k-means, and k-medoids) were compared. K-means clustering performed much better for the simulated data than the other two methods and enabled classification with misclassification error below 5% in the simulated cohort of 100 patients based on the molecular signatures of 40 differentially abundant proteins (effect size 1.5) from among the 1129-protein panel. PMID:27524871
PI2GIS: processing image to geographical information systems, a learning tool for QGIS

NASA Astrophysics Data System (ADS)

Correia, R.; Teodoro, A.; Duarte, L.

2017-10-01

To perform an accurate interpretation of remote sensing images, it is necessary to extract information using different image processing techniques. Nowadays, it became usual to use image processing plugins to add new capabilities/functionalities integrated in Geographical Information System (GIS) software. The aim of this work was to develop an open source application to automatically process and classify remote sensing images from a set of satellite input data. The application was integrated in a GIS software (QGIS), automating several image processing steps. The use of QGIS for this purpose is justified since it is easy and quick to develop new plugins, using Python language. This plugin is inspired in the Semi-Automatic Classification Plugin (SCP) developed by Luca Congedo. SCP allows the supervised classification of remote sensing images, the calculation of vegetation indices such as NDVI (Normalized Difference Vegetation Index) and EVI (Enhanced Vegetation Index) and other image processing operations. When analysing SCP, it was realized that a set of operations, that are very useful in teaching classes of remote sensing and image processing tasks, were lacking, such as the visualization of histograms, the application of filters, different image corrections, unsupervised classification and several environmental indices computation. The new set of operations included in the PI2GIS plugin can be divided into three groups: pre-processing, processing, and classification procedures. The application was tested consider an image from Landsat 8 OLI from a North area of Portugal.
Comparison of Unsupervised Vegetation Classification Methods from Vhr Images after Shadows Removal by Innovative Algorithms

NASA Astrophysics Data System (ADS)

Movia, A.; Beinat, A.; Crosilla, F.

2015-04-01

The recognition of vegetation by the analysis of very high resolution (VHR) aerial images provides meaningful information about environmental features; nevertheless, VHR images frequently contain shadows that generate significant problems for the classification of the image components and for the extraction of the needed information. The aim of this research is to classify, from VHR aerial images, vegetation involved in the balance process of the environmental biochemical cycle, and to discriminate it with respect to urban and agricultural features. Three classification algorithms have been experimented in order to better recognize vegetation, and compared to NDVI index; unfortunately all these methods are conditioned by the presence of shadows on the images. Literature presents several algorithms to detect and remove shadows in the scene: most of them are based on the RGB to HSI transformations. In this work some of them have been implemented and compared with one based on RGB bands. Successively, in order to remove shadows and restore brightness on the images, some innovative algorithms, based on Procrustes theory, have been implemented and applied. Among these, we evaluate the capability of the so called "not-centered oblique Procrustes" and "anisotropic Procrustes" methods to efficiently restore brightness with respect to a linear correlation correction based on the Cholesky decomposition. Some experimental results obtained by different classification methods after shadows removal carried out with the innovative algorithms are presented and discussed.

Unsupervised classification of surface defects in wire rod production obtained by eddy current sensors.

PubMed

Saludes-Rodil, Sergio; Baeyens, Enrique; Rodríguez-Juan, Carlos P

2015-04-29

An unsupervised approach to classify surface defects in wire rod manufacturing is developed in this paper. The defects are extracted from an eddy current signal and classified using a clustering technique that uses the dynamic time warping distance as the dissimilarity measure. The new approach has been successfully tested using industrial data. It is shown that it outperforms other classification alternatives, such as the modified Fourier descriptors.
Weakly supervised visual dictionary learning by harnessing image attributes.

PubMed

Gao, Yue; Ji, Rongrong; Liu, Wei; Dai, Qionghai; Hua, Gang

2014-12-01

Bag-of-features (BoFs) representation has been extensively applied to deal with various computer vision applications. To extract discriminative and descriptive BoF, one important step is to learn a good dictionary to minimize the quantization loss between local features and codewords. While most existing visual dictionary learning approaches are engaged with unsupervised feature quantization, the latest trend has turned to supervised learning by harnessing the semantic labels of images or regions. However, such labels are typically too expensive to acquire, which restricts the scalability of supervised dictionary learning approaches. In this paper, we propose to leverage image attributes to weakly supervise the dictionary learning procedure without requiring any actual labels. As a key contribution, our approach establishes a generative hidden Markov random field (HMRF), which models the quantized codewords as the observed states and the image attributes as the hidden states, respectively. Dictionary learning is then performed by supervised grouping the observed states, where the supervised information is stemmed from the hidden states of the HMRF. In such a way, the proposed dictionary learning approach incorporates the image attributes to learn a semantic-preserving BoF representation without any genuine supervision. Experiments in large-scale image retrieval and classification tasks corroborate that our approach significantly outperforms the state-of-the-art unsupervised dictionary learning approaches.
Aircraft MSS data registration and vegetation classification of wetland change detection

USGS Publications Warehouse

Christensen, E.J.; Jensen, J.R.; Ramsey, Elijah W.; Mackey, H.E.

1988-01-01

Portions of the Savannah River floodplain swamp were evaluated for vegetation change using high resolution (5a??6 m) aircraft multispectral scanner (MSS) data. Image distortion from aircraft movement prevented precise image-to-image registration in some areas. However, when small scenes were used (200-250 ha), a first-order linear transformation provided registration accuracies of less than or equal to one pixel. A larger area was registered using a piecewise linear method. Five major wetland classes were identified and evaluated for change. Phenological differences and the variable distribution of vegetation limited wetland type discrimination. Using unsupervised methods and ground-collected vegetation data, overall classification accuracies ranged from 84 per cent to 87 per cent for each scene. Results suggest that high-resolution aircraft MSS data can be precisely registered, if small areas are used, and that wetland vegetation change can be accurately detected and monitored.
Integration of multispectral satellite and hyperspectral field data for aquatic macrophyte studies

NASA Astrophysics Data System (ADS)

John, C. M.; Kavya, N.

2014-11-01

Aquatic macrophytes (AM) can serve as useful indicators of water pollution along the littoral zones. The spectral signatures of various AM were investigated to determine whether species could be discriminated by remote sensing. In this study the spectral readings of different AM communities identified were done using the ASD Fieldspec® Hand Held spectro-radiometer in the wavelength range of 325-1075 nm. The collected specific reflectance spectra were applied to space borne multi-spectral remote sensing data from Worldview-2, acquired on 26th March 2011. The dimensionality reduction of the spectro-radiometric data was done using the technique principal components analysis (PCA). Out of the different PCA axes generated, 93.472 % variance of the spectra was explained by the first axis. The spectral derivative analysis was done to identify the wavelength where the greatest difference in reflectance is shown. The identified wavelengths are 510, 690, 720, 756, 806, 885, 907 and 923 nm. The output of PCA and derivative analysis were applied to Worldview-2 satellite data for spectral subsetting. The unsupervised classification was used to effectively classify the AM species using the different spectral subsets. The accuracy assessment of the results of the unsupervised classification and their comparison were done. The overall accuracy of the result of unsupervised classification using the band combinations Red-Edge, Green, Coastal blue & Red-edge, Yellow, Blue is 100%. The band combinations NIR-1, Green, Coastal blue & NIR-1, Yellow, Blue yielded an accuracy of 82.35 %. The existing vegetation indices and new hyper-spectral indices for the different type of AM communities were computed. Overall, results of this study suggest that high spectral and spatial resolution images provide useful information for natural resource managers especially with regard to the location identification and distribution mapping of macrophyte species and their communities.
Comparison Between Spectral, Spatial and Polarimetric Classification of Urban and Periurban Landcover Using Temporal Sentinel - 1 Images

NASA Astrophysics Data System (ADS)

Roychowdhury, K.

2016-06-01

Landcover is the easiest detectable indicator of human interventions on land. Urban and peri-urban areas present a complex combination of landcover, which makes classification challenging. This paper assesses the different methods of classifying landcover using dual polarimetric Sentinel-1 data collected during monsoon (July) and winter (December) months of 2015. Four broad landcover classes such as built up areas, water bodies and wetlands, vegetation and open spaces of Kolkata and its surrounding regions were identified. Polarimetric analyses were conducted on Single Look Complex (SLC) data of the region while ground range detected (GRD) data were used for spectral and spatial classification. Unsupervised classification by means of K-Means clustering used backscatter values and was able to identify homogenous landcovers over the study area. The results produced an overall accuracy of less than 50% for both the seasons. Higher classification accuracy (around 70%) was achieved by adding texture variables as inputs along with the backscatter values. However, the accuracy of classification increased significantly with polarimetric analyses. The overall accuracy was around 80% in Wishart H-A-Alpha unsupervised classification. The method was useful in identifying urban areas due to their double-bounce scattering and vegetated areas, which have more random scattering. Normalized Difference Built-up index (NDBI) and Normalized Difference Vegetation Index (NDVI) obtained from Landsat 8 data over the study area were used to verify vegetation and urban classes. The study compares the accuracies of different methods of classifying landcover using medium resolution SAR data in a complex urban area and suggests that polarimetric analyses present the most accurate results for urban and suburban areas.
Quantifying tree mortality in a mixed species woodland using multitemporal high spatial resolution satellite imagery

USGS Publications Warehouse

Garrity, Steven R.; Allen, Craig D.; Brumby, Steven P.; Gangodagamage, Chandana; McDowell, Nate G.; Cai, D. Michael

2013-01-01

Widespread tree mortality events have recently been observed in several biomes. To effectively quantify the severity and extent of these events, tools that allow for rapid assessment at the landscape scale are required. Past studies using high spatial resolution satellite imagery have primarily focused on detecting green, red, and gray tree canopies during and shortly after tree damage or mortality has occurred. However, detecting trees in various stages of death is not always possible due to limited availability of archived satellite imagery. Here we assess the capability of high spatial resolution satellite imagery for tree mortality detection in a southwestern U.S. mixed species woodland using archived satellite images acquired prior to mortality and well after dead trees had dropped their leaves. We developed a multistep classification approach that uses: supervised masking of non-tree image elements; bi-temporal (pre- and post-mortality) differencing of normalized difference vegetation index (NDVI) and red:green ratio (RGI); and unsupervised multivariate clustering of pixels into live and dead tree classes using a Gaussian mixture model. Classification accuracies were improved in a final step by tuning the rules of pixel classification using the posterior probabilities of class membership obtained from the Gaussian mixture model. Classifications were produced for two images acquired post-mortality with overall accuracies of 97.9% and 98.5%, respectively. Classified images were combined with land cover data to characterize the spatiotemporal characteristics of tree mortality across areas with differences in tree species composition. We found that 38% of tree crown area was lost during the drought period between 2002 and 2006. The majority of tree mortality during this period was concentrated in piñon-juniper (Pinus edulis-Juniperus monosperma) woodlands. An additional 20% of the tree canopy died or was removed between 2006 and 2011, primarily in areas experiencing wildfire and management activity. -Our results demonstrate that unsupervised clustering of bi-temporal NDVI and RGI differences can be used to detect tree mortality resulting from numerous causes and in several forest cover types.
An Example of Unsupervised Networks Kohonen's Self-Organizing Feature Map

NASA Technical Reports Server (NTRS)

Niebur, Dagmar

1995-01-01

Kohonen's self-organizing feature map belongs to a class of unsupervised artificial neural network commonly referred to as topographic maps. It serves two purposes, the quantization and dimensionality reduction of date. A short description of its history and its biological context is given. We show that the inherent classification properties of the feature map make it a suitable candidate for solving the classification task in power system areas like load forecasting, fault diagnosis and security assessment.
Cases of Coastal Zone Change and Land Use/Land Cover Change: a learning module that goes beyond the "how" of doing image processing and change detection to asking the "why" about what are the "driving forces" of global change.

NASA Astrophysics Data System (ADS)

Ford, R. E.

2006-12-01

In 2006 the Loma Linda University ESSE21 Mesoamerican Project (Earth System Science Education for the 21st Century) along with partners such as the University of Redlands and California State University, Pomona, produced an online learning module that is designed to help students learn critical remote sensing skills-- specifically: ecosystem characterization, i.e. doing a supervised or unsupervised classification of satellite imagery in a tropical coastal environment. And, it would teach how to measure land use / land cover change (LULC) over time and then encourage students to use that data to assess the Human Dimensions of Global Change (HDGC). Specific objectives include: 1. Learn where to find remote sensing data and practice downloading, pre-processing, and "cleaning" the data for image analysis. 2. Use Leica-Geosystems ERDAS Imagine or IDRISI Kilimanjaro to analyze and display the data. 3. Do an unsupervised classification of a LANDSAT image of a protected area in Honduras, i.e. Cuero y Salado, Pico Bonito, or Isla del Tigre. 4. Virtually participate in a ground-validation exercise that would allow one to re-classify the image into a supervised classification using the FAO Global Land Cover Network (GLCN) classification system. 5. Learn more about each protected area's landscape, history, livelihood patterns and "sustainability" issues via virtual online tours that provide ground and space photos of different sites. This will help students in identifying potential "training sites" for doing a supervised classification. 6. Study other global, US, Canadian, and European land use/land cover classification systems and compare their advantages and disadvantages over the FAO/GLCN system. 7. Learn to appreciate the advantages and disadvantages of existing LULC classification schemes and adapt them to local-level user needs. 8. Carry out a change detection exercise that shows how land use and/or land cover has changed over time for the protected area of your choice. The presenter will demonstrate the module, assess the collaborative process which created it, and describe how it has been used so far by users in the US as well as in Honduras and elsewhere via a series joint workshops held in Mesoamerica. Suggestions for improvement will be requested. See the module and related content resources at: http://resweb.llu.edu/rford/ESSE21/LUCCModule/
Analysis of simulated advanced spaceborne thermal emission and reflection (ASTER) radiometer data of the Iron Hill, Colorado, study area for mapping lithologies

USGS Publications Warehouse

Rowan, L.C.

1998-01-01

The advanced spaceborne thermal emission and reflection (ASTER) radiometer was designed to record reflected energy in nine channels with 15 or 30 m resolution, including stereoscopic images, and emitted energy in five channels with 90 m resolution from the NASA Earth Observing System AM1 platform. A simulated ASTER data set was produced for the Iron Hill, Colorado, study area by resampling calibrated, registered airborne visible/infrared imaging spectrometer (AVIRIS) data, and thermal infrared multispectral scanner (TIMS) data to the appropriate spatial and spectral parameters. A digital elevation model was obtained to simulate ASTER-derived topographic data. The main lithologic units in the area are granitic rocks and felsite into which a carbonatite stock and associated alkalic igneous rocks were intruded; these rocks are locally covered by Jurassic sandstone, Tertiary rhyolitic tuff, and colluvial deposits. Several methods were evaluated for mapping the main lithologic units, including the unsupervised classification and spectral curve-matching techniques. In the five thermal-infrared (TIR) channels, comparison of the results of linear spectral unmixing and unsupervised classification with published geologic maps showed that the main lithologic units were mapped, but large areas with moderate to dense tree cover were not mapped in the TIR data. Compared to TIMS data, simulated ASTER data permitted slightly less discrimination in the mafic alkalic rock series, and carbonatite was not mapped in the TIMS nor in the simulated ASTER TIR data. In the nine visible and near-infrared channels, unsupervised classification did not yield useful results, but both the spectral linear unmixing and the matched filter techniques produced useful results, including mapping calcitic and dolomitic carbonatite exposures, travertine in hot spring deposits, kaolinite in argillized sandstone and tuff, and muscovite in sericitized granite and felsite, as well as commonly occurring illite/muscovite. However, the distinction made in AVIRIS data between calcite and dolomite was not consistently feasible in the simulated ASTER data. Comparison of the lithologic information produced by spectral analysis of the simulated ASTER data to a photogeologic interpretation of a simulated ASTER color image illustrates the high potential of spectral analysis of ASTER data to geologic interpretation. This paper is not subject to U.S. copyright. Published in 1998 by the American Geophysical Union.
Unsupervised classification of operator workload from brain signals.

PubMed

Schultze-Kraft, Matthias; Dähne, Sven; Gugler, Manfred; Curio, Gabriel; Blankertz, Benjamin

2016-06-01

In this study we aimed for the classification of operator workload as it is expected in many real-life workplace environments. We explored brain-signal based workload predictors that differ with respect to the level of label information required for training, including entirely unsupervised approaches. Subjects executed a task on a touch screen that required continuous effort of visual and motor processing with alternating difficulty. We first employed classical approaches for workload state classification that operate on the sensor space of EEG and compared those to the performance of three state-of-the-art spatial filtering methods: common spatial patterns (CSPs) analysis, which requires binary label information; source power co-modulation (SPoC) analysis, which uses the subjects' error rate as a target function; and canonical SPoC (cSPoC) analysis, which solely makes use of cross-frequency power correlations induced by different states of workload and thus represents an unsupervised approach. Finally, we investigated the effects of fusing brain signals and peripheral physiological measures (PPMs) and examined the added value for improving classification performance. Mean classification accuracies of 94%, 92% and 82% were achieved with CSP, SPoC, cSPoC, respectively. These methods outperformed the approaches that did not use spatial filtering and they extracted physiologically plausible components. The performance of the unsupervised cSPoC is significantly increased by augmenting it with PPM features. Our analyses ensured that the signal sources used for classification were of cortical origin and not contaminated with artifacts. Our findings show that workload states can be successfully differentiated from brain signals, even when less and less information from the experimental paradigm is used, thus paving the way for real-world applications in which label information may be noisy or entirely unavailable.
Unsupervised classification of operator workload from brain signals

NASA Astrophysics Data System (ADS)

Schultze-Kraft, Matthias; Dähne, Sven; Gugler, Manfred; Curio, Gabriel; Blankertz, Benjamin

2016-06-01

Objective. In this study we aimed for the classification of operator workload as it is expected in many real-life workplace environments. We explored brain-signal based workload predictors that differ with respect to the level of label information required for training, including entirely unsupervised approaches. Approach. Subjects executed a task on a touch screen that required continuous effort of visual and motor processing with alternating difficulty. We first employed classical approaches for workload state classification that operate on the sensor space of EEG and compared those to the performance of three state-of-the-art spatial filtering methods: common spatial patterns (CSPs) analysis, which requires binary label information; source power co-modulation (SPoC) analysis, which uses the subjects’ error rate as a target function; and canonical SPoC (cSPoC) analysis, which solely makes use of cross-frequency power correlations induced by different states of workload and thus represents an unsupervised approach. Finally, we investigated the effects of fusing brain signals and peripheral physiological measures (PPMs) and examined the added value for improving classification performance. Main results. Mean classification accuracies of 94%, 92% and 82% were achieved with CSP, SPoC, cSPoC, respectively. These methods outperformed the approaches that did not use spatial filtering and they extracted physiologically plausible components. The performance of the unsupervised cSPoC is significantly increased by augmenting it with PPM features. Significance. Our analyses ensured that the signal sources used for classification were of cortical origin and not contaminated with artifacts. Our findings show that workload states can be successfully differentiated from brain signals, even when less and less information from the experimental paradigm is used, thus paving the way for real-world applications in which label information may be noisy or entirely unavailable.
Evaluation of the effects of the seasonal variation of solar elevation angle and azimuth on the processes of digital filtering and thematic classification of relief units

NASA Technical Reports Server (NTRS)

Parada, N. D. J. (Principal Investigator); Novo, E. M. L. M.

1983-01-01

The effects of the seasonal variation of illumination over digital processing of LANDSAT images are evaluated. Two sets of LANDSAT data referring to the orbit 150 and row 28 were selected with illumination parameters varying from 43 deg to 64 deg for azimuth and from 30 deg to 36 deg for solar elevation respectively. IMAGE-100 system permitted the digital processing of LANDSAT data. Original images were transformed by means of digital filtering so as to enhance their spatial features. The resulting images were used to obtain an unsupervised classification of relief units. Topographic variables (declivity, altitude, relief range and slope length) were used to identify the true relief units existing on the ground. The LANDSAT over pass data show that digital processing is highly affected by illumination geometry, and there is no correspondence between relief units as defined by spectral features and those resulting from topographic features.
Segmentation methodology for automated classification and differentiation of soft tissues in multiband images of high-resolution ultrasonic transmission tomography.

PubMed

Jeong, Jeong-Won; Shin, Dae C; Do, Synho; Marmarelis, Vasilis Z

2006-08-01

This paper presents a novel segmentation methodology for automated classification and differentiation of soft tissues using multiband data obtained with the newly developed system of high-resolution ultrasonic transmission tomography (HUTT) for imaging biological organs. This methodology extends and combines two existing approaches: the L-level set active contour (AC) segmentation approach and the agglomerative hierarchical kappa-means approach for unsupervised clustering (UC). To prevent the trapping of the current iterative minimization AC algorithm in a local minimum, we introduce a multiresolution approach that applies the level set functions at successively increasing resolutions of the image data. The resulting AC clusters are subsequently rearranged by the UC algorithm that seeks the optimal set of clusters yielding the minimum within-cluster distances in the feature space. The presented results from Monte Carlo simulations and experimental animal-tissue data demonstrate that the proposed methodology outperforms other existing methods without depending on heuristic parameters and provides a reliable means for soft tissue differentiation in HUTT images.
Model-Based Building Detection from Low-Cost Optical Sensors Onboard Unmanned Aerial Vehicles

NASA Astrophysics Data System (ADS)

Karantzalos, K.; Koutsourakis, P.; Kalisperakis, I.; Grammatikopoulos, L.

2015-08-01

The automated and cost-effective building detection in ultra high spatial resolution is of major importance for various engineering and smart city applications. To this end, in this paper, a model-based building detection technique has been developed able to extract and reconstruct buildings from UAV aerial imagery and low-cost imaging sensors. In particular, the developed approach through advanced structure from motion, bundle adjustment and dense image matching computes a DSM and a true orthomosaic from the numerous GoPro images which are characterised by important geometric distortions and fish-eye effect. An unsupervised multi-region, graphcut segmentation and a rule-based classification is responsible for delivering the initial multi-class classification map. The DTM is then calculated based on inpaininting and mathematical morphology process. A data fusion process between the detected building from the DSM/DTM and the classification map feeds a grammar-based building reconstruction and scene building are extracted and reconstructed. Preliminary experimental results appear quite promising with the quantitative evaluation indicating detection rates at object level of 88% regarding the correctness and above 75% regarding the detection completeness.
Area Series Handbook: Hungary: A Country Study

DTIC Science & Technology

1991-02-01

Wi. c hmeier indl Smith mote iniager, interpi-t.nton. and (JS5 systems. 1 978). Because R) is generally constant within an area Readers interested in...MSS) image (#850871 16251 X0, 6 July 20, 1986) was obtained for the Fort Hood area. A Products computer- generated , unsupervised classification wasmade...conservation prac- reciprocal of C, the equation becomes tices, however, are generally incompatible with military EI = (R x K x LS x P)/T training. Therefore
Unsupervised classification of scattering behavior using radar polarimetry data

NASA Technical Reports Server (NTRS)

Van Zyl, Jakob J.

1989-01-01

The use of an imaging radar polarimeter data for unsupervised classification of scattering behavior is described by comparing the polarization properties of each pixel in a image to that of simple classes of scattering such as even number of reflections, odd number of reflections, and diffuse scattering. For example, when this algorithm is applied to data acquired over the San Francisco Bay area in California, it classifies scattering by the ocean as being similar to that predicted by the class of odd number of reflections, scattering by the urban area as being similar to that predicted by the class of even number of reflections, and scattering by the Golden Gate Park as being similar to that predicted by the diffuse scattering class. It also classifies the scattering by a lighthouse in the ocean and boats on the ocean surface as being similar to that predicted by the even number of reflections class, making it easy to identify these objects against the background of the surrounding ocean. The algorithm is also applied to forested areas and shows that scattering from clear-cut areas and agricultural fields is mostly similar to that predicted by the odd number of reflections class, while the scattering from tree-covered areas generally is classified as being a mixture of pixels exhibiting the characteristics of all three classes, although each pixel is identified with only a single class.
Wide field imaging - I. Applications of neural networks to object detection and star/galaxy classification

NASA Astrophysics Data System (ADS)

Andreon, S.; Gargiulo, G.; Longo, G.; Tagliaferri, R.; Capuano, N.

2000-12-01

Astronomical wide-field imaging performed with new large-format CCD detectors poses data reduction problems of unprecedented scale, which are difficult to deal with using traditional interactive tools. We present here NExt (Neural Extractor), a new neural network (NN) based package capable of detecting objects and performing both deblending and star/galaxy classification in an automatic way. Traditionally, in astronomical images, objects are first distinguished from the noisy background by searching for sets of connected pixels having brightnesses above a given threshold; they are then classified as stars or as galaxies through diagnostic diagrams having variables chosen according to the astronomer's taste and experience. In the extraction step, assuming that images are well sampled, NExt requires only the simplest a priori definition of `what an object is' (i.e. it keeps all structures composed of more than one pixel) and performs the detection via an unsupervised NN, approaching detection as a clustering problem that has been thoroughly studied in the artificial intelligence literature. The first part of the NExt procedure consists of an optimal compression of the redundant information contained in the pixels via a mapping from pixel intensities to a subspace individualized through principal component analysis. At magnitudes fainter than the completeness limit, stars are usually almost indistinguishable from galaxies, and therefore the parameters characterizing the two classes do not lie in disconnected subspaces, thus preventing the use of unsupervised methods. We therefore adopted a supervised NN (i.e. a NN that first finds the rules to classify objects from examples and then applies them to the whole data set). In practice, each object is classified depending on its membership of the regions mapping the input feature space in the training set. In order to obtain an objective and reliable classification, instead of using an arbitrarily defined set of features we use a NN to select the most significant features among the large number of measured ones, and then we use these selected features to perform the classification task. In order to optimize the performance of the system, we implemented and tested several different models of NN. The comparison of the NExt performance with that of the best detection and classification package known to the authors (SExtractor) shows that NExt is at least as effective as the best traditional packages.
Classification of multispectral image data by the Binary Diamond neural network and by nonparametric, pixel-by-pixel methods

NASA Technical Reports Server (NTRS)

Salu, Yehuda; Tilton, James

1993-01-01

The classification of multispectral image data obtained from satellites has become an important tool for generating ground cover maps. This study deals with the application of nonparametric pixel-by-pixel classification methods in the classification of pixels, based on their multispectral data. A new neural network, the Binary Diamond, is introduced, and its performance is compared with a nearest neighbor algorithm and a back-propagation network. The Binary Diamond is a multilayer, feed-forward neural network, which learns from examples in unsupervised, 'one-shot' mode. It recruits its neurons according to the actual training set, as it learns. The comparisons of the algorithms were done by using a realistic data base, consisting of approximately 90,000 Landsat 4 Thematic Mapper pixels. The Binary Diamond and the nearest neighbor performances were close, with some advantages to the Binary Diamond. The performance of the back-propagation network lagged behind. An efficient nearest neighbor algorithm, the binned nearest neighbor, is described. Ways for improving the performances, such as merging categories, and analyzing nonboundary pixels, are addressed and evaluated.
Analysis of the Tanana River Basin using LANDSAT data

NASA Technical Reports Server (NTRS)

Morrissey, L. A.; Ambrosia, V. G.; Carson-Henry, C.

1981-01-01

Digital image classification techniques were used to classify land cover/resource information in the Tanana River Basin of Alaska. Portions of four scenes of LANDSAT digital data were analyzed using computer systems at Ames Research Center in an unsupervised approach to derive cluster statistics. The spectral classes were identified using the IDIMS display and color infrared photography. Classification errors were corrected using stratification procedures. The classification scheme resulted in the following eleven categories; sedimented/shallow water, clear/deep water, coniferous forest, mixed forest, deciduous forest, shrub and grass, bog, alpine tundra, barrens, snow and ice, and cultural features. Color coded maps and acreage summaries of the major land cover categories were generated for selected USGS quadrangles (1:250,000) which lie within the drainage basin. The project was completed within six months.
Classification of leafy spurge with earth observing-1 advanced land imager

USGS Publications Warehouse

Stitt, S.; Root, R.; Brown, K.; Hager, S.; Mladinich, C.; Anderson, G.L.; Dudek, K.; Bustos, M.R.; Kokaly, R.

2006-01-01

Leafy spurge (Euphorbia esula L.) is an invasive exotic plant that can completely displace native plant communities. Automated techniques for monitoring the location and extent of leafy spurge, especially if available on a seasonal basis, could add greatly to the effectiveness of control measures. As part of a larger study including multiple sensors, this study examines the utility of mapping the location and extent of leafy spurge in Theodore Roosevelt National Park using Earth Observing-1 satellite Advanced Land Imager (ALI) scanner data. An unsupervised classification methodology was used producing accuracies in the range of 59% to 66%. Existing field studies, with their associated limitations, were used for identifying class membership and accuracy assessment. This sensor could be useful for broad landscape scale mapping of leafy spurge, from which control measures could be based.

Physical Human Activity Recognition Using Wearable Sensors.

PubMed

Attal, Ferhat; Mohammed, Samer; Dedabrishvili, Mariam; Chamroukhi, Faicel; Oukhellou, Latifa; Amirat, Yacine

2015-12-11

This paper presents a review of different classification techniques used to recognize human activities from wearable inertial sensor data. Three inertial sensor units were used in this study and were worn by healthy subjects at key points of upper/lower body limbs (chest, right thigh and left ankle). Three main steps describe the activity recognition process: sensors' placement, data pre-processing and data classification. Four supervised classification techniques namely, k-Nearest Neighbor (k-NN), Support Vector Machines (SVM), Gaussian Mixture Models (GMM), and Random Forest (RF) as well as three unsupervised classification techniques namely, k-Means, Gaussian mixture models (GMM) and Hidden Markov Model (HMM), are compared in terms of correct classification rate, F-measure, recall, precision, and specificity. Raw data and extracted features are used separately as inputs of each classifier. The feature selection is performed using a wrapper approach based on the RF algorithm. Based on our experiments, the results obtained show that the k-NN classifier provides the best performance compared to other supervised classification algorithms, whereas the HMM classifier is the one that gives the best results among unsupervised classification algorithms. This comparison highlights which approach gives better performance in both supervised and unsupervised contexts. It should be noted that the obtained results are limited to the context of this study, which concerns the classification of the main daily living human activities using three wearable accelerometers placed at the chest, right shank and left ankle of the subject.
Physical Human Activity Recognition Using Wearable Sensors

PubMed Central

Attal, Ferhat; Mohammed, Samer; Dedabrishvili, Mariam; Chamroukhi, Faicel; Oukhellou, Latifa; Amirat, Yacine

2015-01-01

This paper presents a review of different classification techniques used to recognize human activities from wearable inertial sensor data. Three inertial sensor units were used in this study and were worn by healthy subjects at key points of upper/lower body limbs (chest, right thigh and left ankle). Three main steps describe the activity recognition process: sensors’ placement, data pre-processing and data classification. Four supervised classification techniques namely, k-Nearest Neighbor (k-NN), Support Vector Machines (SVM), Gaussian Mixture Models (GMM), and Random Forest (RF) as well as three unsupervised classification techniques namely, k-Means, Gaussian mixture models (GMM) and Hidden Markov Model (HMM), are compared in terms of correct classification rate, F-measure, recall, precision, and specificity. Raw data and extracted features are used separately as inputs of each classifier. The feature selection is performed using a wrapper approach based on the RF algorithm. Based on our experiments, the results obtained show that the k-NN classifier provides the best performance compared to other supervised classification algorithms, whereas the HMM classifier is the one that gives the best results among unsupervised classification algorithms. This comparison highlights which approach gives better performance in both supervised and unsupervised contexts. It should be noted that the obtained results are limited to the context of this study, which concerns the classification of the main daily living human activities using three wearable accelerometers placed at the chest, right shank and left ankle of the subject. PMID:26690450
Image Classification Workflow Using Machine Learning Methods

NASA Astrophysics Data System (ADS)

Christoffersen, M. S.; Roser, M.; Valadez-Vergara, R.; Fernández-Vega, J. A.; Pierce, S. A.; Arora, R.

2016-12-01

Recent increases in the availability and quality of remote sensing datasets have fueled an increasing number of scientifically significant discoveries based on land use classification and land use change analysis. However, much of the software made to work with remote sensing data products, specifically multispectral images, is commercial and often prohibitively expensive. The free to use solutions that are currently available come bundled up as small parts of much larger programs that are very susceptible to bugs and difficult to install and configure. What is needed is a compact, easy to use set of tools to perform land use analysis on multispectral images. To address this need, we have developed software using the Python programming language with the sole function of land use classification and land use change analysis. We chose Python to develop our software because it is relatively readable, has a large body of relevant third party libraries such as GDAL and Spectral Python, and is free to install and use on Windows, Linux, and Macintosh operating systems. In order to test our classification software, we performed a K-means unsupervised classification, Gaussian Maximum Likelihood supervised classification, and a Mahalanobis Distance based supervised classification. The images used for testing were three Landsat rasters of Austin, Texas with a spatial resolution of 60 meters for the years of 1984 and 1999, and 30 meters for the year 2015. The testing dataset was easily downloaded using the Earth Explorer application produced by the USGS. The software should be able to perform classification based on any set of multispectral rasters with little to no modification. Our software makes the ease of land use classification using commercial software available without an expensive license.
Semi-supervised and unsupervised extreme learning machines.

PubMed

Huang, Gao; Song, Shiji; Gupta, Jatinder N D; Wu, Cheng

2014-12-01

Extreme learning machines (ELMs) have proven to be efficient and effective learning mechanisms for pattern classification and regression. However, ELMs are primarily applied to supervised learning problems. Only a few existing research papers have used ELMs to explore unlabeled data. In this paper, we extend ELMs for both semi-supervised and unsupervised tasks based on the manifold regularization, thus greatly expanding the applicability of ELMs. The key advantages of the proposed algorithms are as follows: 1) both the semi-supervised ELM (SS-ELM) and the unsupervised ELM (US-ELM) exhibit learning capability and computational efficiency of ELMs; 2) both algorithms naturally handle multiclass classification or multicluster clustering; and 3) both algorithms are inductive and can handle unseen data at test time directly. Moreover, it is shown in this paper that all the supervised, semi-supervised, and unsupervised ELMs can actually be put into a unified framework. This provides new perspectives for understanding the mechanism of random feature mapping, which is the key concept in ELM theory. Empirical study on a wide range of data sets demonstrates that the proposed algorithms are competitive with the state-of-the-art semi-supervised or unsupervised learning algorithms in terms of accuracy and efficiency.
Impervious surface mapping with Quickbird imagery

PubMed Central

Lu, Dengsheng; Hetrick, Scott; Moran, Emilio

2010-01-01

This research selects two study areas with different urban developments, sizes, and spatial patterns to explore the suitable methods for mapping impervious surface distribution using Quickbird imagery. The selected methods include per-pixel based supervised classification, segmentation-based classification, and a hybrid method. A comparative analysis of the results indicates that per-pixel based supervised classification produces a large number of “salt-and-pepper” pixels, and segmentation based methods can significantly reduce this problem. However, neither method can effectively solve the spectral confusion of impervious surfaces with water/wetland and bare soils and the impacts of shadows. In order to accurately map impervious surface distribution from Quickbird images, manual editing is necessary and may be the only way to extract impervious surfaces from the confused land covers and the shadow problem. This research indicates that the hybrid method consisting of thresholding techniques, unsupervised classification and limited manual editing provides the best performance. PMID:21643434
Unsupervised tattoo segmentation combining bottom-up and top-down cues

NASA Astrophysics Data System (ADS)

Allen, Josef D.; Zhao, Nan; Yuan, Jiangbo; Liu, Xiuwen

2011-06-01

Tattoo segmentation is challenging due to the complexity and large variance in tattoo structures. We have developed a segmentation algorithm for finding tattoos in an image. Our basic idea is split-merge: split each tattoo image into clusters through a bottom-up process, learn to merge the clusters containing skin and then distinguish tattoo from the other skin via top-down prior in the image itself. Tattoo segmentation with unknown number of clusters is transferred to a figureground segmentation. We have applied our segmentation algorithm on a tattoo dataset and the results have shown that our tattoo segmentation system is efficient and suitable for further tattoo classification and retrieval purpose.
Alzheimer disease detection from structural MR images using FCM based weighted probabilistic neural network.

PubMed

Duraisamy, Baskar; Shanmugam, Jayanthi Venkatraman; Annamalai, Jayanthi

2018-02-19

An early intervention of Alzheimer's disease (AD) is highly essential due to the fact that this neuro degenerative disease generates major life-threatening issues, especially memory loss among patients in society. Moreover, categorizing NC (Normal Control), MCI (Mild Cognitive Impairment) and AD early in course allows the patients to experience benefits from new treatments. Therefore, it is important to construct a reliable classification technique to discriminate the patients with or without AD from the bio medical imaging modality. Hence, we developed a novel FCM based Weighted Probabilistic Neural Network (FWPNN) classification algorithm and analyzed the brain images related to structural MRI modality for better discrimination of class labels. Initially our proposed framework begins with brain image normalization stage. In this stage, ROI regions related to Hippo-Campus (HC) and Posterior Cingulate Cortex (PCC) from the brain images are extracted using Automated Anatomical Labeling (AAL) method. Subsequently, nineteen highly relevant AD related features are selected through Multiple-criterion feature selection method. At last, our novel FWPNN classification algorithm is imposed to remove suspicious samples from the training data with an end goal to enhance the classification performance. This newly developed classification algorithm combines both the goodness of supervised and unsupervised learning techniques. The experimental validation is carried out with the ADNI subset and then to the Bordex-3 city dataset. Our proposed classification approach achieves an accuracy of about 98.63%, 95.4%, 96.4% in terms of classification with AD vs NC, MCI vs NC and AD vs MCI. The experimental results suggest that the removal of noisy samples from the training data can enhance the decision generation process of the expert systems.
Adaptive fuzzy leader clustering of complex data sets in pattern recognition

NASA Technical Reports Server (NTRS)

Newton, Scott C.; Pemmaraju, Surya; Mitra, Sunanda

1992-01-01

A modular, unsupervised neural network architecture for clustering and classification of complex data sets is presented. The adaptive fuzzy leader clustering (AFLC) architecture is a hybrid neural-fuzzy system that learns on-line in a stable and efficient manner. The initial classification is performed in two stages: a simple competitive stage and a distance metric comparison stage. The cluster prototypes are then incrementally updated by relocating the centroid positions from fuzzy C-means system equations for the centroids and the membership values. The AFLC algorithm is applied to the Anderson Iris data and laser-luminescent fingerprint image data. It is concluded that the AFLC algorithm successfully classifies features extracted from real data, discrete or continuous.
Linear mixing model applied to coarse spatial resolution data from multispectral satellite sensors

NASA Technical Reports Server (NTRS)

Holben, Brent N.; Shimabukuro, Yosio E.

1993-01-01

A linear mixing model was applied to coarse spatial resolution data from the NOAA Advanced Very High Resolution Radiometer. The reflective component of the 3.55-3.95 micron channel was used with the two reflective channels 0.58-0.68 micron and 0.725-1.1 micron to run a constrained least squares model to generate fraction images for an area in the west central region of Brazil. The fraction images were compared with an unsupervised classification derived from Landsat TM data acquired on the same day. The relationship between the fraction images and normalized difference vegetation index images show the potential of the unmixing techniques when using coarse spatial resolution data for global studies.
Automatic cloud coverage assessment of Formosat-2 image

NASA Astrophysics Data System (ADS)

Hsu, Kuo-Hsien

2011-11-01

Formosat-2 satellite equips with the high-spatial-resolution (2m ground sampling distance) remote sensing instrument. It has been being operated on the daily-revisiting mission orbit by National Space organization (NSPO) of Taiwan since May 21 2004. NSPO has also serving as one of the ground receiving stations for daily processing the received Formosat- 2 images. The current cloud coverage assessment of Formosat-2 image for NSPO Image Processing System generally consists of two major steps. Firstly, an un-supervised K-means method is used for automatically estimating the cloud statistic of Formosat-2 image. Secondly, manual estimation of cloud coverage from Formosat-2 image is processed by manual examination. Apparently, a more accurate Automatic Cloud Coverage Assessment (ACCA) method certainly increases the efficiency of processing step 2 with a good prediction of cloud statistic. In this paper, mainly based on the research results from Chang et al, Irish, and Gotoh, we propose a modified Formosat-2 ACCA method which considered pre-processing and post-processing analysis. For pre-processing analysis, cloud statistic is determined by using un-supervised K-means classification, Sobel's method, Otsu's method, non-cloudy pixels reexamination, and cross-band filter method. Box-Counting fractal method is considered as a post-processing tool to double check the results of pre-processing analysis for increasing the efficiency of manual examination.
Classify epithelium-stroma in histopathological images based on deep transferable network.

PubMed

Yu, X; Zheng, H; Liu, C; Huang, Y; Ding, X

2018-04-20

Recently, the deep learning methods have received more attention in histopathological image analysis. However, the traditional deep learning methods assume that training data and test data have the same distributions, which causes certain limitations in real-world histopathological applications. However, it is costly to recollect a large amount of labeled histology data to train a new neural network for each specified image acquisition procedure even for similar tasks. In this paper, an unsupervised domain adaptation is introduced into a typical deep convolutional neural network (CNN) model to mitigate the repeating of the labels. The unsupervised domain adaptation is implemented by adding two regularisation terms, namely the feature-based adaptation and entropy minimisation, to the object function of a widely used CNN model called the AlexNet. Three independent public epithelium-stroma datasets were used to verify the proposed method. The experimental results have demonstrated that in the epithelium-stroma classification, the proposed method can achieve better performance than the commonly used deep learning methods and some existing deep domain adaptation methods. Therefore, the proposed method can be considered as a better option for the real-world applications of histopathological image analysis because there is no requirement for recollection of large-scale labeled data for every specified domain. © 2018 The Authors Journal of Microscopy © 2018 Royal Microscopical Society.
Colour space influence for vegetation image classification application to Caribbean forest and agriculture

NASA Astrophysics Data System (ADS)

Abadi, M.; Grandchamp, E.

2008-10-01

This paper deals with a comparison of different colour space in order to improve high resolution images classification. The background of this study is the measure of the agriculture impact on the environment in islander context. Biodiversity is particularly sensitive and relevant in such areas and the follow-up of the forest front is a way to ensure its preservation. Very high resolution satellite images are used such as QuickBird and IKONOS scenes. In order to segment the images into forest and agriculture areas, we characterize both ground covers with colour and texture features. A classical unsupervised classifier is then used to obtain labelled areas. As features are computed on coloured images, we can wonder if the colour space choice is relevant. This study has been made considering more than fourteen colour spaces (RGB, YUV, Lab, YIQ, YCrCs, XYZ, CMY, LMS, HSL, KLT, IHS, I1I2I3, HSV, HSI, etc.) and shows the visual and quantitative superiority of IHS on all others. For conciseness reasons, results only show RGB, I1I2I3 and IHS colour spaces.
Feature Extraction Using an Unsupervised Neural Network

DTIC Science & Technology

1991-05-03

with this neural netowrk is given and its connection to exploratory projection pursuit methods is established. DD I 2 P JA d 73 EDITIONj Of I NOV 6s...IS OBSOLETE $IN 0102- LF- 014- 6601 SECURITY CLASSIFICATION OF THIS PAGE (When Daoes Enlered) Feature Extraction using an Unsupervised Neural Network
Predicting protein complexes using a supervised learning method combined with local structural information.

PubMed

Dong, Yadong; Sun, Yongqi; Qin, Chao

2018-01-01

The existing protein complex detection methods can be broadly divided into two categories: unsupervised and supervised learning methods. Most of the unsupervised learning methods assume that protein complexes are in dense regions of protein-protein interaction (PPI) networks even though many true complexes are not dense subgraphs. Supervised learning methods utilize the informative properties of known complexes; they often extract features from existing complexes and then use the features to train a classification model. The trained model is used to guide the search process for new complexes. However, insufficient extracted features, noise in the PPI data and the incompleteness of complex data make the classification model imprecise. Consequently, the classification model is not sufficient for guiding the detection of complexes. Therefore, we propose a new robust score function that combines the classification model with local structural information. Based on the score function, we provide a search method that works both forwards and backwards. The results from experiments on six benchmark PPI datasets and three protein complex datasets show that our approach can achieve better performance compared with the state-of-the-art supervised, semi-supervised and unsupervised methods for protein complex detection, occasionally significantly outperforming such methods.
The composite sequential clustering technique for analysis of multispectral scanner data

NASA Technical Reports Server (NTRS)

Su, M. Y.

1972-01-01

The clustering technique consists of two parts: (1) a sequential statistical clustering which is essentially a sequential variance analysis, and (2) a generalized K-means clustering. In this composite clustering technique, the output of (1) is a set of initial clusters which are input to (2) for further improvement by an iterative scheme. This unsupervised composite technique was employed for automatic classification of two sets of remote multispectral earth resource observations. The classification accuracy by the unsupervised technique is found to be comparable to that by traditional supervised maximum likelihood classification techniques. The mathematical algorithms for the composite sequential clustering program and a detailed computer program description with job setup are given.
Identification of sea ice types in spaceborne synthetic aperture radar data

NASA Technical Reports Server (NTRS)

Kwok, Ronald; Rignot, Eric; Holt, Benjamin; Onstott, R.

1992-01-01

This study presents an approach for identification of sea ice types in spaceborne SAR image data. The unsupervised classification approach involves cluster analysis for segmentation of the image data followed by cluster labeling based on previously defined look-up tables containing the expected backscatter signatures of different ice types measured by a land-based scatterometer. Extensive scatterometer observations and experience accumulated in field campaigns during the last 10 yr were used to construct these look-up tables. The classification approach, its expected performance, the dependence of this performance on radar system performance, and expected ice scattering characteristics are discussed. Results using both aircraft and simulated ERS-1 SAR data are presented and compared to limited field ice property measurements and coincident passive microwave imagery. The importance of an integrated postlaunch program for the validation and improvement of this approach is discussed.
Principal component analysis-based unsupervised feature extraction applied to in silico drug discovery for posttraumatic stress disorder-mediated heart disease.

PubMed

Taguchi, Y-h; Iwadate, Mitsuo; Umeyama, Hideaki

2015-04-30

Feature extraction (FE) is difficult, particularly if there are more features than samples, as small sample numbers often result in biased outcomes or overfitting. Furthermore, multiple sample classes often complicate FE because evaluating performance, which is usual in supervised FE, is generally harder than the two-class problem. Developing sample classification independent unsupervised methods would solve many of these problems. Two principal component analysis (PCA)-based FE, specifically, variational Bayes PCA (VBPCA) was extended to perform unsupervised FE, and together with conventional PCA (CPCA)-based unsupervised FE, were tested as sample classification independent unsupervised FE methods. VBPCA- and CPCA-based unsupervised FE both performed well when applied to simulated data, and a posttraumatic stress disorder (PTSD)-mediated heart disease data set that had multiple categorical class observations in mRNA/microRNA expression of stressed mouse heart. A critical set of PTSD miRNAs/mRNAs were identified that show aberrant expression between treatment and control samples, and significant, negative correlation with one another. Moreover, greater stability and biological feasibility than conventional supervised FE was also demonstrated. Based on the results obtained, in silico drug discovery was performed as translational validation of the methods. Our two proposed unsupervised FE methods (CPCA- and VBPCA-based) worked well on simulated data, and outperformed two conventional supervised FE methods on a real data set. Thus, these two methods have suggested equivalence for FE on categorical multiclass data sets, with potential translational utility for in silico drug discovery.
Hyperspectral and Hypertemporal Longwave Infrared Data Characterization

NASA Astrophysics Data System (ADS)

Jeganathan, Nirmalan

The Army Research Lab conducted a persistent imaging experiment called the Spectral and Polarimetric Imagery Collection Experiment (SPICE) in 2012 and 2013 which focused on collecting and exploiting long wave infrared hyperspectral and polarimetric imagery. A part of this dataset was made for public release for research and development purposes. This thesis investigated the hyperspectral portion of this released dataset through data characterization and scene characterization of man-made and natural objects. First, the data were contrasted with MODerate resolution atmospheric TRANsmission (MODTRAN) results and found to be comparable. Instrument noise was characterized using an in-scene black panel, and was found to be comparable with the sensor manufacturer's specication. The temporal and spatial variation of certain objects in the scene were characterized. Temporal target detection was conducted on man-made objects in the scene using three target detection algorithms: spectral angle mapper (SAM), spectral matched lter (SMF) and adaptive coherence/cosine estimator (ACE). SMF produced the best results for detecting the targets when the training and testing data originated from different time periods, with a time index percentage result of 52.9%. Unsupervised and supervised classification were conducted using spectral and temporal target signatures. Temporal target signatures produced better visual classification than spectral target signature for unsupervised classification. Supervised classification yielded better results using the spectral target signatures, with a highest weighted accuracy of 99% for 7-class reference image. Four emissivity retrieval algorithms were applied on this dataset. However, the retrieved emissivities from all four methods did not represent true material emissivity and could not be used for analysis. This spectrally and temporally rich dataset enabled to conduct analysis that was not possible with other data collections. Regarding future work, applying noise-reduction techniques before applying temperature-emissivity retrieval algorithms may produce more realistic emissivity values, which could be used for target detection and material identification.
Development of remote sensing based site specific weed management for Midwest mint production

NASA Astrophysics Data System (ADS)

Gumz, Mary Saumur Paulson

Peppermint and spearmint are high value essential oil crops in Indiana, Michigan, and Wisconsin. Although the mints are profitable alternatives to corn and soybeans, mint production efficiency must improve in order to allow industry survival against foreign produced oils and synthetic flavorings. Weed control is the major input cost in mint production and tools to increase efficiency are necessary. Remote sensing-based site-specific weed management offers potential for decreasing weed control costs through simplified weed detection and control from accurate site specific weed and herbicide application maps. This research showed the practicability of remote sensing for weed detection in the mints. Research was designed to compare spectral response curves of field grown mint and weeds, and to use these data to develop spectral vegetation indices for automated weed detection. Viability of remote sensing in mint production was established using unsupervised classification, supervised classification, handheld spectroradiometer readings and spectral vegetation indices (SVIs). Unsupervised classification of multispectral images of peppermint production fields generated crop health maps with 92 and 67% accuracy in meadow and row peppermint, respectively. Supervised classification of multispectral images identified weed infestations with 97% and 85% accuracy for meadow and row peppermint, respectively. Supervised classification showed that peppermint was spectrally distinct from weeds, but the accuracy of these measures was dependent on extensive ground referencing which is impractical and too costly for on-farm use. Handheld spectroradiometer measurements of peppermint, spearmint, and several weeds and crop and weed mixtures were taken over three years from greenhouse grown plants, replicated field plots, and production peppermint and spearmint fields. Results showed that mints have greater near infrared (NIR) and lower green reflectance and a steeper red edge slope than all weed species. These distinguishing characteristics were combined to develop narrow band and broadband spectral vegetation indices (SVIs, ratios of NIR/green reflectance), that were effective in differentiating mint from key weed species. Hyperspectral images of production peppermint and spearmint fields were then classified using SVI-based classification. Narrowband and broadband SVIs classified early season peppermint and spearmint with 64 to 100% accuracy compared to 79 to 100% accuracy for supervised classification of multispectral images of the same fields. Broadband SVIs have potential for use as an automated spectral indicator for weeds in the mints since they require minimal ground referencing and can be calculated from multispectral imagery which is cheaper and more readily available than hyperspectral imagery. This research will allow growers to implement remote sensing based site specific weed management in mint resulting in reduced grower input costs and reduced herbicide entry into the environment and will have applications in other specialty and meadow crops.
Can segmentation evaluation metric be used as an indicator of land cover classification accuracy?

NASA Astrophysics Data System (ADS)

Švab Lenarčič, Andreja; Đurić, Nataša; Čotar, Klemen; Ritlop, Klemen; Oštir, Krištof

2016-10-01

It is a broadly established belief that the segmentation result significantly affects subsequent image classification accuracy. However, the actual correlation between the two has never been evaluated. Such an evaluation would be of considerable importance for any attempts to automate the object-based classification process, as it would reduce the amount of user intervention required to fine-tune the segmentation parameters. We conducted an assessment of segmentation and classification by analyzing 100 different segmentation parameter combinations, 3 classifiers, 5 land cover classes, 20 segmentation evaluation metrics, and 7 classification accuracy measures. The reliability definition of segmentation evaluation metrics as indicators of land cover classification accuracy was based on the linear correlation between the two. All unsupervised metrics that are not based on number of segments have a very strong correlation with all classification measures and are therefore reliable as indicators of land cover classification accuracy. On the other hand, correlation at supervised metrics is dependent on so many factors that it cannot be trusted as a reliable classification quality indicator. Algorithms for land cover classification studied in this paper are widely used; therefore, presented results are applicable to a wider area.

Single-particle cryo-EM using alignment by classification (ABC): the structure of Lumbricus terrestris haemoglobin.

PubMed

Afanasyev, Pavel; Seer-Linnemayr, Charlotte; Ravelli, Raimond B G; Matadeen, Rishi; De Carlo, Sacha; Alewijnse, Bart; Portugal, Rodrigo V; Pannu, Navraj S; Schatz, Michael; van Heel, Marin

2017-09-01

Single-particle cryogenic electron microscopy (cryo-EM) can now yield near-atomic resolution structures of biological complexes. However, the reference-based alignment algorithms commonly used in cryo-EM suffer from reference bias, limiting their applicability (also known as the 'Einstein from random noise' problem). Low-dose cryo-EM therefore requires robust and objective approaches to reveal the structural information contained in the extremely noisy data, especially when dealing with small structures. A reference-free pipeline is presented for obtaining near-atomic resolution three-dimensional reconstructions from heterogeneous ('four-dimensional') cryo-EM data sets. The methodologies integrated in this pipeline include a posteriori camera correction, movie-based full-data-set contrast transfer function determination, movie-alignment algorithms, (Fourier-space) multivariate statistical data compression and unsupervised classification, 'random-startup' three-dimensional reconstructions, four-dimensional structural refinements and Fourier shell correlation criteria for evaluating anisotropic resolution. The procedures exclusively use information emerging from the data set itself, without external 'starting models'. Euler-angle assignments are performed by angular reconstitution rather than by the inherently slower projection-matching approaches. The comprehensive 'ABC-4D' pipeline is based on the two-dimensional reference-free 'alignment by classification' (ABC) approach, where similar images in similar orientations are grouped by unsupervised classification. Some fundamental differences between X-ray crystallography versus single-particle cryo-EM data collection and data processing are discussed. The structure of the giant haemoglobin from Lumbricus terrestris at a global resolution of ∼3.8 Å is presented as an example of the use of the ABC-4D procedure.
A Novel Unsupervised Adaptive Learning Method for Long-Term Electromyography (EMG) Pattern Recognition

PubMed Central

Huang, Qi; Yang, Dapeng; Jiang, Li; Zhang, Huajie; Liu, Hong; Kotani, Kiyoshi

2017-01-01

Performance degradation will be caused by a variety of interfering factors for pattern recognition-based myoelectric control methods in the long term. This paper proposes an adaptive learning method with low computational cost to mitigate the effect in unsupervised adaptive learning scenarios. We presents a particle adaptive classifier (PAC), by constructing a particle adaptive learning strategy and universal incremental least square support vector classifier (LS-SVC). We compared PAC performance with incremental support vector classifier (ISVC) and non-adapting SVC (NSVC) in a long-term pattern recognition task in both unsupervised and supervised adaptive learning scenarios. Retraining time cost and recognition accuracy were compared by validating the classification performance on both simulated and realistic long-term EMG data. The classification results of realistic long-term EMG data showed that the PAC significantly decreased the performance degradation in unsupervised adaptive learning scenarios compared with NSVC (9.03% ± 2.23%, p < 0.05) and ISVC (13.38% ± 2.62%, p = 0.001), and reduced the retraining time cost compared with ISVC (2 ms per updating cycle vs. 50 ms per updating cycle). PMID:28608824
A Novel Unsupervised Adaptive Learning Method for Long-Term Electromyography (EMG) Pattern Recognition.

PubMed

Huang, Qi; Yang, Dapeng; Jiang, Li; Zhang, Huajie; Liu, Hong; Kotani, Kiyoshi

2017-06-13

Performance degradation will be caused by a variety of interfering factors for pattern recognition-based myoelectric control methods in the long term. This paper proposes an adaptive learning method with low computational cost to mitigate the effect in unsupervised adaptive learning scenarios. We presents a particle adaptive classifier (PAC), by constructing a particle adaptive learning strategy and universal incremental least square support vector classifier (LS-SVC). We compared PAC performance with incremental support vector classifier (ISVC) and non-adapting SVC (NSVC) in a long-term pattern recognition task in both unsupervised and supervised adaptive learning scenarios. Retraining time cost and recognition accuracy were compared by validating the classification performance on both simulated and realistic long-term EMG data. The classification results of realistic long-term EMG data showed that the PAC significantly decreased the performance degradation in unsupervised adaptive learning scenarios compared with NSVC (9.03% ± 2.23%, p < 0.05) and ISVC (13.38% ± 2.62%, p = 0.001), and reduced the retraining time cost compared with ISVC (2 ms per updating cycle vs. 50 ms per updating cycle).
Remote sensing of Earth terrain

NASA Technical Reports Server (NTRS)

Kong, Jin AU; Shin, Robert T.; Nghiem, Son V.; Yueh, Herng-Aung; Han, Hsiu C.; Lim, Harold H.; Arnold, David V.

1990-01-01

Remote sensing of earth terrain is examined. The layered random medium model is used to investigate the fully polarimetric scattering of electromagnetic waves from vegetation. The model is used to interpret the measured data for vegetation fields such as rice, wheat, or soybean over water or soil. Accurate calibration of polarimetric radar systems is essential for the polarimetric remote sensing of earth terrain. A polarimetric calibration algorithm using three arbitrary in-scene reflectors is developed. In the interpretation of active and passive microwave remote sensing data from the earth terrain, the random medium model was shown to be quite successful. A multivariate K-distribution is proposed to model the statistics of fully polarimetric radar returns from earth terrain. In the terrain cover classification using the synthetic aperture radar (SAR) images, the applications of the K-distribution model will provide better performance than the conventional Gaussian classifiers. The layered random medium model is used to study the polarimetric response of sea ice. Supervised and unsupervised classification procedures are also developed and applied to synthetic aperture radar polarimetric images in order to identify their various earth terrain components for more than two classes. These classification procedures were applied to San Francisco Bay and Traverse City SAR images.
PixelLearn

NASA Technical Reports Server (NTRS)

Mazzoni, Dominic; Wagstaff, Kiri; Bornstein, Benjamin; Tang, Nghia; Roden, Joseph

2006-01-01

PixelLearn is an integrated user-interface computer program for classifying pixels in scientific images. Heretofore, training a machine-learning algorithm to classify pixels in images has been tedious and difficult. PixelLearn provides a graphical user interface that makes it faster and more intuitive, leading to more interactive exploration of image data sets. PixelLearn also provides image-enhancement controls to make it easier to see subtle details in images. PixelLearn opens images or sets of images in a variety of common scientific file formats and enables the user to interact with several supervised or unsupervised machine-learning pixel-classifying algorithms while the user continues to browse through the images. The machinelearning algorithms in PixelLearn use advanced clustering and classification methods that enable accuracy much higher than is achievable by most other software previously available for this purpose. PixelLearn is written in portable C++ and runs natively on computers running Linux, Windows, or Mac OS X.
Remote sensing of submerged aquatic vegetation in lower Chesapeake Bay - A comparison of Landsat MSS to TM imagery

NASA Technical Reports Server (NTRS)

Ackleson, S. G.; Klemas, V.

1987-01-01

Landsat MSS and TM imagery, obtained simultaneously over Guinea Marsh, VA, as analyzed and compares for its ability to detect submerged aquatic vegetation (SAV). An unsupervised clustering algorithm was applied to each image, where the input classification parameters are defined as functions of apparent sensor noise. Class confidence and accuracy were computed for all water areas by comparing the classified images, pixel-by-pixel, to rasterized SAV distributions derived from color aerial photography. To illustrate the effect of water depth on classification error, areas of depth greater than 1.9 m were masked, and class confidence and accuracy recalculated. A single-scattering radiative-transfer model is used to illustrate how percent canopy cover and water depth affect the volume reflectance from a water column containing SAV. For a submerged canopy that is morphologically and optically similar to Zostera marina inhabiting Lower Chesapeake Bay, dense canopies may be isolated by masking optically deep water. For less dense canopies, the effect of increasing water depth is to increase the apparent percent crown cover, which may result in classification error.
A simple and unsupervised semi-automatic workflow to detect shallow landslides in Alpine areas based on VHR remote sensing data

NASA Astrophysics Data System (ADS)

Amato, Gabriele; Eisank, Clemens; Albrecht, Florian

2017-04-01

Landslide detection from Earth observation imagery is an important preliminary work for landslide mapping, landslide inventories and landslide hazard assessment. In this context, the object-based image analysis (OBIA) concept has been increasingly used over the last decade. Within the framework of the Land@Slide project (Earth observation based landslide mapping: from methodological developments to automated web-based information delivery) a simple, unsupervised, semi-automatic and object-based approach for the detection of shallow landslides has been developed and implemented in the InterIMAGE open-source software. The method was applied to an Alpine case study in western Austria, exploiting spectral information from pansharpened 4-bands WorldView-2 satellite imagery (0.5 m spatial resolution) in combination with digital elevation models. First, we divided the image into sub-images, i.e. tiles, and then we applied the workflow to each of them without changing the parameters. The workflow was implemented as top-down approach: at the image tile level, an over-classification of the potential landslide area was produced; the over-estimated area was re-segmented and re-classified by several processing cycles until most false positive objects have been eliminated. In every step a Baatz algorithm based segmentation generates polygons "candidates" to be landslides. At the same time, the average values of normalized difference vegetation index (NDVI) and brightness are calculated for these polygons; after that, these values are used as thresholds to perform an objects selection in order to improve the quality of the classification results. In combination, also empirically determined values of slope and roughness are used in the selection process. Results for each tile were merged to obtain the landslide map for the test area. For final validation, the landslide map was compared to a geological map and a supervised landslide classification in order to estimate its accuracy. Results for the test area showed that the proposed method is capable of accurately distinguishing landslides from roofs and trees. Implementation of the workflow into InterIMAGE was straightforward. We conclude that the method is able to extract landslides in forested areas, but that there is still room for improvements concerning the extraction in non-forested high-alpine regions.
Colorectal cancer detection by hyperspectral imaging using fluorescence excitation scanning

NASA Astrophysics Data System (ADS)

Leavesley, Silas J.; Deal, Joshua; Hill, Shante; Martin, Will A.; Lall, Malvika; Lopez, Carmen; Rider, Paul F.; Rich, Thomas C.; Boudreaux, Carole W.

2018-02-01

Hyperspectral imaging technologies have shown great promise for biomedical applications. These techniques have been especially useful for detection of molecular events and characterization of cell, tissue, and biomaterial composition. Unfortunately, hyperspectral imaging technologies have been slow to translate to clinical devices - likely due to increased cost and complexity of the technology as well as long acquisition times often required to sample a spectral image. We have demonstrated that hyperspectral imaging approaches which scan the fluorescence excitation spectrum can provide increased signal strength and faster imaging, compared to traditional emission-scanning approaches. We have also demonstrated that excitation-scanning approaches may be able to detect spectral differences between colonic adenomas and adenocarcinomas and normal mucosa in flash-frozen tissues. Here, we report feasibility results from using excitation-scanning hyperspectral imaging to screen pairs of fresh tumoral and nontumoral colorectal tissues. Tissues were imaged using a novel hyperspectral imaging fluorescence excitation scanning microscope, sampling a wavelength range of 360-550 nm, at 5 nm increments. Image data were corrected to achieve a NIST-traceable flat spectral response. Image data were then analyzed using a range of supervised and unsupervised classification approaches within ENVI software (Harris Geospatial Solutions). Supervised classification resulted in >99% accuracy for single-patient image data, but only 64% accuracy for multi-patient classification (n=9 to date), with the drop in accuracy due to increased false-positive detection rates. Hence, initial data indicate that this approach may be a viable detection approach, but that larger patient sample sizes need to be evaluated and the effects of inter-patient variability studied.
Bayesian Fusion of Color and Texture Segmentations

NASA Technical Reports Server (NTRS)

Manduchi, Roberto

2000-01-01

In many applications one would like to use information from both color and texture features in order to segment an image. We propose a novel technique to combine "soft" segmentations computed for two or more features independently. Our algorithm merges models according to a mean entropy criterion, and allows to choose the appropriate number of classes for the final grouping. This technique also allows to improve the quality of supervised classification based on one feature (e.g. color) by merging information from unsupervised segmentation based on another feature (e.g., texture.)
Evaluating the Visualization of What a Deep Neural Network Has Learned.

PubMed

Samek, Wojciech; Binder, Alexander; Montavon, Gregoire; Lapuschkin, Sebastian; Muller, Klaus-Robert

Deep neural networks (DNNs) have demonstrated impressive performance in complex machine learning tasks such as image classification or speech recognition. However, due to their multilayer nonlinear structure, they are not transparent, i.e., it is hard to grasp what makes them arrive at a particular classification or recognition decision, given a new unseen data sample. Recently, several approaches have been proposed enabling one to understand and interpret the reasoning embodied in a DNN for a single test image. These methods quantify the "importance" of individual pixels with respect to the classification decision and allow a visualization in terms of a heatmap in pixel/input space. While the usefulness of heatmaps can be judged subjectively by a human, an objective quality measure is missing. In this paper, we present a general methodology based on region perturbation for evaluating ordered collections of pixels such as heatmaps. We compare heatmaps computed by three different methods on the SUN397, ILSVRC2012, and MIT Places data sets. Our main result is that the recently proposed layer-wise relevance propagation algorithm qualitatively and quantitatively provides a better explanation of what made a DNN arrive at a particular classification decision than the sensitivity-based approach or the deconvolution method. We provide theoretical arguments to explain this result and discuss its practical implications. Finally, we investigate the use of heatmaps for unsupervised assessment of the neural network performance.Deep neural networks (DNNs) have demonstrated impressive performance in complex machine learning tasks such as image classification or speech recognition. However, due to their multilayer nonlinear structure, they are not transparent, i.e., it is hard to grasp what makes them arrive at a particular classification or recognition decision, given a new unseen data sample. Recently, several approaches have been proposed enabling one to understand and interpret the reasoning embodied in a DNN for a single test image. These methods quantify the "importance" of individual pixels with respect to the classification decision and allow a visualization in terms of a heatmap in pixel/input space. While the usefulness of heatmaps can be judged subjectively by a human, an objective quality measure is missing. In this paper, we present a general methodology based on region perturbation for evaluating ordered collections of pixels such as heatmaps. We compare heatmaps computed by three different methods on the SUN397, ILSVRC2012, and MIT Places data sets. Our main result is that the recently proposed layer-wise relevance propagation algorithm qualitatively and quantitatively provides a better explanation of what made a DNN arrive at a particular classification decision than the sensitivity-based approach or the deconvolution method. We provide theoretical arguments to explain this result and discuss its practical implications. Finally, we investigate the use of heatmaps for unsupervised assessment of the neural network performance.
Ensemble Semi-supervised Frame-work for Brain Magnetic Resonance Imaging Tissue Segmentation.

PubMed

Azmi, Reza; Pishgoo, Boshra; Norozi, Narges; Yeganeh, Samira

2013-04-01

Brain magnetic resonance images (MRIs) tissue segmentation is one of the most important parts of the clinical diagnostic tools. Pixel classification methods have been frequently used in the image segmentation with two supervised and unsupervised approaches up to now. Supervised segmentation methods lead to high accuracy, but they need a large amount of labeled data, which is hard, expensive, and slow to obtain. Moreover, they cannot use unlabeled data to train classifiers. On the other hand, unsupervised segmentation methods have no prior knowledge and lead to low level of performance. However, semi-supervised learning which uses a few labeled data together with a large amount of unlabeled data causes higher accuracy with less trouble. In this paper, we propose an ensemble semi-supervised frame-work for segmenting of brain magnetic resonance imaging (MRI) tissues that it has been used results of several semi-supervised classifiers simultaneously. Selecting appropriate classifiers has a significant role in the performance of this frame-work. Hence, in this paper, we present two semi-supervised algorithms expectation filtering maximization and MCo_Training that are improved versions of semi-supervised methods expectation maximization and Co_Training and increase segmentation accuracy. Afterward, we use these improved classifiers together with graph-based semi-supervised classifier as components of the ensemble frame-work. Experimental results show that performance of segmentation in this approach is higher than both supervised methods and the individual semi-supervised classifiers.
Feature Selection for Ridge Regression with Provable Guarantees.

PubMed

Paul, Saurabh; Drineas, Petros

2016-04-01

We introduce single-set spectral sparsification as a deterministic sampling-based feature selection technique for regularized least-squares classification, which is the classification analog to ridge regression. The method is unsupervised and gives worst-case guarantees of the generalization power of the classification function after feature selection with respect to the classification function obtained using all features. We also introduce leverage-score sampling as an unsupervised randomized feature selection method for ridge regression. We provide risk bounds for both single-set spectral sparsification and leverage-score sampling on ridge regression in the fixed design setting and show that the risk in the sampled space is comparable to the risk in the full-feature space. We perform experiments on synthetic and real-world data sets; a subset of TechTC-300 data sets, to support our theory. Experimental results indicate that the proposed methods perform better than the existing feature selection methods.
Fossil Signatures Using Elemental Abundance Distributions and Bayesian Probabilistic Classification

NASA Technical Reports Server (NTRS)

Hoover, Richard B.; Storrie-Lombardi, Michael C.

2004-01-01

Elemental abundances (C6, N7, O8, Na11, Mg12, Al3, P15, S16, Cl17, K19, Ca20, Ti22, Mn25, Fe26, and Ni28) were obtained for a set of terrestrial fossils and the rock matrix surrounding them. Principal Component Analysis extracted five factors accounting for the 92.5% of the data variance, i.e. information content, of the elemental abundance data. Hierarchical Cluster Analysis provided unsupervised sample classification distinguishing fossil from matrix samples on the basis of either raw abundances or PCA input that agreed strongly with visual classification. A stochastic, non-linear Artificial Neural Network produced a Bayesian probability of correct sample classification. The results provide a quantitative probabilistic methodology for discriminating terrestrial fossils from the surrounding rock matrix using chemical information. To demonstrate the applicability of these techniques to the assessment of meteoritic samples or in situ extraterrestrial exploration, we present preliminary data on samples of the Orgueil meteorite. In both systems an elemental signature produces target classification decisions remarkably consistent with morphological classification by a human expert using only structural (visual) information. We discuss the possibility of implementing a complexity analysis metric capable of automating certain image analysis and pattern recognition abilities of the human eye using low magnification optical microscopy images and discuss the extension of this technique across multiple scales.
Learned filters for object detection in multi-object visual tracking

NASA Astrophysics Data System (ADS)

Stamatescu, Victor; Wong, Sebastien; McDonnell, Mark D.; Kearney, David

2016-05-01

We investigate the application of learned convolutional filters in multi-object visual tracking. The filters were learned in both a supervised and unsupervised manner from image data using artificial neural networks. This work follows recent results in the field of machine learning that demonstrate the use learned filters for enhanced object detection and classification. Here we employ a track-before-detect approach to multi-object tracking, where tracking guides the detection process. The object detection provides a probabilistic input image calculated by selecting from features obtained using banks of generative or discriminative learned filters. We present a systematic evaluation of these convolutional filters using a real-world data set that examines their performance as generic object detectors.
Improving zero-training brain-computer interfaces by mixing model estimators

NASA Astrophysics Data System (ADS)

Verhoeven, T.; Hübner, D.; Tangermann, M.; Müller, K. R.; Dambre, J.; Kindermans, P. J.

2017-06-01

Objective. Brain-computer interfaces (BCI) based on event-related potentials (ERP) incorporate a decoder to classify recorded brain signals and subsequently select a control signal that drives a computer application. Standard supervised BCI decoders require a tedious calibration procedure prior to every session. Several unsupervised classification methods have been proposed that tune the decoder during actual use and as such omit this calibration. Each of these methods has its own strengths and weaknesses. Our aim is to improve overall accuracy of ERP-based BCIs without calibration. Approach. We consider two approaches for unsupervised classification of ERP signals. Learning from label proportions (LLP) was recently shown to be guaranteed to converge to a supervised decoder when enough data is available. In contrast, the formerly proposed expectation maximization (EM) based decoding for ERP-BCI does not have this guarantee. However, while this decoder has high variance due to random initialization of its parameters, it obtains a higher accuracy faster than LLP when the initialization is good. We introduce a method to optimally combine these two unsupervised decoding methods, letting one method’s strengths compensate for the weaknesses of the other and vice versa. The new method is compared to the aforementioned methods in a resimulation of an experiment with a visual speller. Main results. Analysis of the experimental results shows that the new method exceeds the performance of the previous unsupervised classification approaches in terms of ERP classification accuracy and symbol selection accuracy during the spelling experiment. Furthermore, the method shows less dependency on random initialization of model parameters and is consequently more reliable. Significance. Improving the accuracy and subsequent reliability of calibrationless BCIs makes these systems more appealing for frequent use.
Diagnostic index of three-dimensional osteoarthritic changes in temporomandibular joint condylar morphology

PubMed Central

Gomes, Liliane R.; Gomes, Marcelo; Jung, Bryan; Paniagua, Beatriz; Ruellas, Antonio C.; Gonçalves, João Roberto; Styner, Martin A.; Wolford, Larry; Cevidanes, Lucia

2015-01-01

Abstract. This study aimed to investigate imaging statistical approaches for classifying three-dimensional (3-D) osteoarthritic morphological variations among 169 temporomandibular joint (TMJ) condyles. Cone-beam computed tomography scans were acquired from 69 subjects with long-term TMJ osteoarthritis (OA), 15 subjects at initial diagnosis of OA, and 7 healthy controls. Three-dimensional surface models of the condyles were constructed and SPHARM-PDM established correspondent points on each model. Multivariate analysis of covariance and direction-projection-permutation (DiProPerm) were used for testing statistical significance of the differences between the groups determined by clinical and radiographic diagnoses. Unsupervised classification using hierarchical agglomerative clustering was then conducted. Compared with healthy controls, OA average condyle was significantly smaller in all dimensions except its anterior surface. Significant flattening of the lateral pole was noticed at initial diagnosis. We observed areas of 3.88-mm bone resorption at the superior surface and 3.10-mm bone apposition at the anterior aspect of the long-term OA average model. DiProPerm supported a significant difference between the healthy control and OA group (p-value=0.001). Clinically meaningful unsupervised classification of TMJ condylar morphology determined a preliminary diagnostic index of 3-D osteoarthritic changes, which may be the first step towards a more targeted diagnosis of this condition. PMID:26158119
Mapping Neglected Swimming Pools from Satellite Data for Urban Vector Control

NASA Astrophysics Data System (ADS)

Barker, C. M.; Melton, F. S.; Reisen, W. K.

2010-12-01

Neglected swimming pools provide suitable breeding habit for mosquitoes, can contain thousands of mosquito larvae, and present both a significant nuisance and public health risk due to their inherent proximity to urban and suburban populations. The rapid increase and sustained rate of foreclosures in California associated with the recent recession presents a challenge for vector control districts seeking to identify, treat, and monitor neglected pools. Commercial high resolution satellite imagery offers some promise for mapping potential neglected pools, and for mapping pools for which routine maintenance has been reestablished. We present progress on unsupervised classification techniques for mapping both neglected pools and clean pools using high resolution commercial satellite data and discuss the potential uses and limitations of this data source in support of vector control efforts. An unsupervised classification scheme that utilizes image segmentation, band thresholds, and a change detection approach was implemented for sample regions in Coachella Valley, CA and the greater Los Angeles area. Comparison with field data collected by vector control personal was used to assess the accuracy of the estimates. The results suggest that the current system may provide some utility for early detection, or cost effective and time efficient annual monitoring, but additional work is required to address spectral and spatial limitations of current commercial satellite sensors for this purpose.
Linear mixing model applied to coarse resolution satellite data

NASA Technical Reports Server (NTRS)

Holben, Brent N.; Shimabukuro, Yosio E.

1992-01-01

A linear mixing model typically applied to high resolution data such as Airborne Visible/Infrared Imaging Spectrometer, Thematic Mapper, and Multispectral Scanner System is applied to the NOAA Advanced Very High Resolution Radiometer coarse resolution satellite data. The reflective portion extracted from the middle IR channel 3 (3.55 - 3.93 microns) is used with channels 1 (0.58 - 0.68 microns) and 2 (0.725 - 1.1 microns) to run the Constrained Least Squares model to generate fraction images for an area in the west central region of Brazil. The derived fraction images are compared with an unsupervised classification and the fraction images derived from Landsat TM data acquired in the same day. In addition, the relationship betweeen these fraction images and the well known NDVI images are presented. The results show the great potential of the unmixing techniques for applying to coarse resolution data for global studies.
An evaluation of unsupervised and supervised learning algorithms for clustering landscape types in the United States

USGS Publications Warehouse

Wendel, Jochen; Buttenfield, Barbara P.; Stanislawski, Larry V.

2016-01-01

Knowledge of landscape type can inform cartographic generalization of hydrographic features, because landscape characteristics provide an important geographic context that affects variation in channel geometry, flow pattern, and network configuration. Landscape types are characterized by expansive spatial gradients, lacking abrupt changes between adjacent classes; and as having a limited number of outliers that might confound classification. The US Geological Survey (USGS) is exploring methods to automate generalization of features in the National Hydrography Data set (NHD), to associate specific sequences of processing operations and parameters with specific landscape characteristics, thus obviating manual selection of a unique processing strategy for every NHD watershed unit. A chronology of methods to delineate physiographic regions for the United States is described, including a recent maximum likelihood classification based on seven input variables. This research compares unsupervised and supervised algorithms applied to these seven input variables, to evaluate and possibly refine the recent classification. Evaluation metrics for unsupervised methods include the Davies–Bouldin index, the Silhouette index, and the Dunn index as well as quantization and topographic error metrics. Cross validation and misclassification rate analysis are used to evaluate supervised classification methods. The paper reports the comparative analysis and its impact on the selection of landscape regions. The compared solutions show problems in areas of high landscape diversity. There is some indication that additional input variables, additional classes, or more sophisticated methods can refine the existing classification.
A SOFTWARE PACKAGE FOR UNSUPERVISED PATTERN RECOGNITION AND SYNOPTIC REPRESENTATION OF RESULTS: APPLICATION TO VOLCANIC TREMOR DATA OF MT ETNA

NASA Astrophysics Data System (ADS)

Langer, H. K.; Falsaperla, S. M.; Behncke, B.; Messina, A.; Spampinato, S.

2009-12-01

Artificial Intelligence (AI) has found broad applications in volcano observatories worldwide with the aim of reducing volcanic hazard. The need to process larger and larger quantity of data makes indeed AI techniques appealing for monitoring purposes. Tools based on Artificial Neural Networks and Support Vector Machine have proved to be particularly successful in the classification of seismic events and volcanic tremor changes heralding eruptive activity, such as paroxysmal explosions and lava fountaining at Stromboli and Mt Etna, Italy (e.g., Falsaperla et al., 1996; Langer et al., 2009). Moving on from the excellent results obtained from these applications, we present KKAnalysis, a MATLAB based software which combines several unsupervised pattern classification methods, exploiting routines of the SOM Toolbox 2 for MATLAB (http://www.cis.hut.fi/projects/somtoolbox). KKAnalysis is based on Self Organizing Maps (SOM) and clustering methods consisting of K-Means, Fuzzy C-Means, and a scheme based on a metrics accounting for correlation between components of the feature vector. We show examples of applications of this tool to volcanic tremor data recorded at Mt Etna between 2007 and 2009. This time span - during which Strombolian explosions, 7 episodes of lava fountaining and effusive activity occurred - is particularly interesting, as it encompassed different states of volcanic activity (i.e., non-eruptive, eruptive according to different styles) for the unsupervised classifier to identify, highlighting their development in time. Even subtle changes in the signal characteristics allow the unsupervised classifier to recognize features belonging to the different classes and stages of volcanic activity. A convenient color-code representation shows up the temporal development of the different classes of signal, making this method extremely helpful for monitoring purposes and surveillance. Though being developed for volcanic tremor classification, KKAnalysis is generally applicable to any type of physical or chemical pattern, provided that feature vectors are given in numerical form. References: Falsaperla, S., S. Graziani, G. Nunnari, and S. Spampinato (1996). Automatic classification of volcanic earthquakes by using multy-layered neural networks. Natural Hazard, 13, 205-228. Langer, H., S. Falsaperla, M. Masotti, R. Campanini, S. Spampinato, and A. Messina (2008). Synopsis of supervised and unsupervised pattern classification techniques applied to volcanic tremor data at Mt Etna, Italy. Geophys. J. Int., doi:10.1111/j.1365-246X.2009.04179.x.

Unsupervised classification of cirrhotic livers using MRI data

NASA Astrophysics Data System (ADS)

Lee, Gobert; Kanematsu, Masayuki; Kato, Hiroki; Kondo, Hiroshi; Zhou, Xiangrong; Hara, Takeshi; Fujita, Hiroshi; Hoshi, Hiroaki

2008-03-01

Cirrhosis of the liver is a chronic disease. It is characterized by the presence of widespread nodules and fibrosis in the liver which results in characteristic texture patterns. Computerized analysis of hepatic texture patterns is usually based on regions-of-interest (ROIs). However, not all ROIs are typical representatives of the disease stage of the liver from which the ROIs originated. This leads to uncertainties in the ROI labels (diseased or non-diseased). On the other hand, supervised classifiers are commonly used in determining the assignment rule. This presents a problem as the training of a supervised classifier requires the correct labels of the ROIs. The main purpose of this paper is to investigate the use of an unsupervised classifier, the k-means clustering, in classifying ROI based data. In addition, a procedure for generating a receiver operating characteristic (ROC) curve depicting the classification performance of k-means clustering is also reported. Hepatic MRI images of 44 patients (16 cirrhotic; 28 non-cirrhotic) are used in this study. The MRI data are derived from gadolinium-enhanced equilibrium phase images. For each patient, 10 ROIs selected by an experienced radiologist and 7 texture features measured on each ROI are included in the MRI data. Results of the k-means classifier are depicted using an ROC curve. The area under the curve (AUC) has a value of 0.704. This is slightly lower than but comparable to that of LDA and ANN classifiers which have values 0.781 and 0.801, respectively. Methods in constructing ROC curve in relation to k-means clustering have not been previously reported in the literature.
SOM Classification of Martian TES Data

NASA Technical Reports Server (NTRS)

Hogan, R. C.; Roush, T. L.

2002-01-01

A classification scheme based on unsupervised self-organizing maps (SOM) is described. Results from its application to the ASU mineral spectral database are presented. Applications to the Martian Thermal Emission Spectrometer data are discussed. Additional information is contained in the original extended abstract.
Remote Sensing Application to Land Use Classification in a Rapidly Changing Agricultural/Urban Area: City of Virginia Beach, Virginia. Ph.D. Thesis

NASA Technical Reports Server (NTRS)

Odenyo, V. A. O.

1975-01-01

Remote sensing data on computer-compatible tapes of LANDSAT 1 multispectral scanner imager were analyzed to generate a land use map of the City of Virginia Beach. All four bands were used in both the supervised and unsupervised approaches with the LAYSYS software system. Color IR imagery of a U-2 flight of the same area was also digitized and two sample areas were analyzed via the unsupervised approach. The relationships between the mapped land use and the soils of the area were investigated. A land use land cover map at a scale of 1:24,000 was obtained from the supervised analysis of LANDSAT 1 data. It was concluded that machine analysis of remote sensing data to produce land use maps was feasible; that the LAYSYS software system was usable for this purpose; and that the machine analysis was capable of extracting detailed information from the relatively small scale LANDSAT data in a much shorter time without compromising accuracy.
Globally maximizing, locally minimizing: unsupervised discriminant projection with applications to face and palm biometrics.

PubMed

Yang, Jian; Zhang, David; Yang, Jing-Yu; Niu, Ben

2007-04-01

This paper develops an unsupervised discriminant projection (UDP) technique for dimensionality reduction of high-dimensional data in small sample size cases. UDP can be seen as a linear approximation of a multimanifolds-based learning framework which takes into account both the local and nonlocal quantities. UDP characterizes the local scatter as well as the nonlocal scatter, seeking to find a projection that simultaneously maximizes the nonlocal scatter and minimizes the local scatter. This characteristic makes UDP more intuitive and more powerful than the most up-to-date method, Locality Preserving Projection (LPP), which considers only the local scatter for clustering or classification tasks. The proposed method is applied to face and palm biometrics and is examined using the Yale, FERET, and AR face image databases and the PolyU palmprint database. The experimental results show that UDP consistently outperforms LPP and PCA and outperforms LDA when the training sample size per class is small. This demonstrates that UDP is a good choice for real-world biometrics applications.
An unsupervised two-stage clustering approach for forest structure classification based on X-band InSAR data - A case study in complex temperate forest stands

NASA Astrophysics Data System (ADS)

Abdullahi, Sahra; Schardt, Mathias; Pretzsch, Hans

2017-05-01

Forest structure at stand level plays a key role for sustainable forest management, since the biodiversity, productivity, growth and stability of the forest can be positively influenced by managing its structural diversity. In contrast to field-based measurements, remote sensing techniques offer a cost-efficient opportunity to collect area-wide information about forest stand structure with high spatial and temporal resolution. Especially Interferometric Synthetic Aperture Radar (InSAR), which facilitates worldwide acquisition of 3d information independent from weather conditions and illumination, is convenient to capture forest stand structure. This study purposes an unsupervised two-stage clustering approach for forest structure classification based on height information derived from interferometric X-band SAR data which was performed in complex temperate forest stands of Traunstein forest (South Germany). In particular, a four dimensional input data set composed of first-order height statistics was non-linearly projected on a two-dimensional Self-Organizing Map, spatially ordered according to similarity (based on the Euclidean distance) in the first stage and classified using the k-means algorithm in the second stage. The study demonstrated that X-band InSAR data exhibits considerable capabilities for forest structure classification. Moreover, the unsupervised classification approach achieved meaningful and reasonable results by means of comparison to aerial imagery and LiDAR data.
Unsupervised domain adaptation for early detection of drought stress in hyperspectral images

NASA Astrophysics Data System (ADS)

Schmitter, P.; Steinrücken, J.; Römer, C.; Ballvora, A.; Léon, J.; Rascher, U.; Plümer, L.

2017-09-01

Hyperspectral images can be used to uncover physiological processes in plants if interpreted properly. Machine Learning methods such as Support Vector Machines (SVM) and Random Forests have been applied to estimate development of biomass and detect and predict plant diseases and drought stress. One basic requirement of machine learning implies, that training and testing is done in the same domain and the same distribution. Different genotypes, environmental conditions, illumination and sensors violate this requirement in most practical circumstances. Here, we present an approach, which enables the detection of physiological processes by transferring the prior knowledge within an existing model into a related target domain, where no label information is available. We propose a two-step transformation of the target features, which enables a direct application of an existing model. The transformation is evaluated by an objective function including additional prior knowledge about classification and physiological processes in plants. We have applied the approach to three sets of hyperspectral images, which were acquired with different plant species in different environments observed with different sensors. It is shown, that a classification model, derived on one of the sets, delivers satisfying classification results on the transformed features of the other data sets. Furthermore, in all cases early non-invasive detection of drought stress was possible.
Techniques for Producing Coastal Land Water Masks from Landsat and Other Multispectral Satellite Data

NASA Technical Reports Server (NTRS)

Spruce, Joseph P.; Hall, Callie

2005-01-01

Coastal erosion and land loss continue to threaten many areas in the United States. Landsat data has been used to monitor regional coastal change since the 1970s. Many techniques can be used to produce coastal land water masks, including image classification and density slicing of individual bands or of band ratios. Band ratios used in land water detection include several variations of the Normalized Difference Water Index (NDWI). This poster discusses a study that compares land water masks computed from unsupervised Landsat image classification with masks from density-sliced band ratios and from the Landsat TM band 5. The greater New Orleans area is employed in this study, due to its abundance of coastal habitats and its vulnerability to coastal land loss. Image classification produced the best results based on visual comparison to higher resolution satellite and aerial image displays. However, density sliced NDWI imagery from either near infrared (NIR) and blue bands or from NIR and green bands also produced more effective land water masks than imagery from the density-sliced Landsat TM band 5. NDWI based on NIR and green bands is noteworthy because it allows land water masks to be generated from multispectral satellite sensors without a blue band (e.g., ASTER and Landsat MSS). NDWI techniques also have potential for producing land water masks from coarser scaled satellite data, such as MODIS.
Techniques for Producing Coastal Land Water Masks from Landsat and Other Multispectral Satellite Data

NASA Technical Reports Server (NTRS)

Spruce, Joe; Hall, Callie

2005-01-01

Coastal erosion and land loss continue to threaten many areas in the United States. Landsat data has been used to monitor regional coastal change since the 1970's. Many techniques can be used to produce coastal land water masks, including image classification and density slicing of individual bands or of band ratios. Band ratios used in land water detection include several variations of the Normalized Difference Water Index (NDWI). This poster discusses a study that compares land water masks computed from unsupervised Landsat image classification with masks from density-sliced band ratios and from the Landsat TM band 5. The greater New Orleans area is imployed in this study, due to its abundance of coastal habitats and ist vulnerability to coastal land loss. Image classification produced the best results based on visual comparison to higher resolution satellite and aerial image displays. However, density-sliced NDWI imagery from either near infrared (NIR) and blue bands or from NIR and green bands also produced more effective land water masks than imagery from the density-sliced Landsat TM band 5. NDWI based on NIR and green bands is noteworthy because it allows land water masks to be generated form multispectral satellite sensors without a blue band (e.g., ASTER and Landsat MSS). NDWI techniques also have potential for producing land water masks from coarser scaled satellite data, such as MODIS.
Landsat TM inventory and assessment of waterbird habitat in the southern altiplano of South America

USGS Publications Warehouse

Boyle, T.P.; Caziani, S.M.; Waltermire, R.G.

2004-01-01

The diverse set of wetlands in southern altiplano of South America supports a number of endemic and migratory waterbirds. These species include endangered endemic flamingos and shorebirds that nest in North America and winter in the altiplano. This research developed maps from nine Landsat Thematic Mapper (TM) images (254,300 km2) to provide an inventory of aquatic waterbird habitats. Image processing software was used to produce a map with a classification of wetlands according to the habitat requirements of different types of waterbirds. A hierarchical procedure was used to, first, isolate the bodies of water within the TM image; second, execute an unsupervised classification on the subsetted image to produce 300 signatures of cover types, which were further subdivided as necessary. Third, each of the classifications was examined in the light of field data and personal experience for relevance to the determination of the various habitat types. Finally, the signatures were applied to the entire image and other adjacent images to yield a map depicting the location of the various waterbird habitats in the southern altiplano. The data sets referenced with a global positioning system receiver were used to test the classification system. Multivariate analysis of the bird communities censused at each lake by individual habitats indicated a salinity gradient, and then the depth of the water separated the birds. Multivariate analysis of the chemical and physical data from the lakes showed that the variation in lakes were significantly associated with difference in depth, transparency, latitude, elevation, and pH. The presence of gravel bottoms was also one of the qualities distinguishing a group of lakes. This information will be directly useful to the Flamingo Census Project and serve as an element for risk assessment for future development.
Classification of grass pollen through the quantitative analysis of surface ornamentation and texture.

PubMed

Mander, Luke; Li, Mao; Mio, Washington; Fowlkes, Charless C; Punyasena, Surangi W

2013-11-07

Taxonomic identification of pollen and spores uses inherently qualitative descriptions of morphology. Consequently, identifications are restricted to categories that can be reliably classified by multiple analysts, resulting in the coarse taxonomic resolution of the pollen and spore record. Grass pollen represents an archetypal example; it is not routinely identified below family level. To address this issue, we developed quantitative morphometric methods to characterize surface ornamentation and classify grass pollen grains. This produces a means of quantifying morphological features that are traditionally described qualitatively. We used scanning electron microscopy to image 240 specimens of pollen from 12 species within the grass family (Poaceae). We classified these species by developing algorithmic features that quantify the size and density of sculptural elements on the pollen surface, and measure the complexity of the ornamentation they form. These features yielded a classification accuracy of 77.5%. In comparison, a texture descriptor based on modelling the statistical distribution of brightness values in image patches yielded a classification accuracy of 85.8%, and seven human subjects achieved accuracies between 68.33 and 81.67%. The algorithmic features we developed directly relate to biologically meaningful features of grass pollen morphology, and could facilitate direct interpretation of unsupervised classification results from fossil material.
Communications and control for electric power systems: Power flow classification for static security assessment

NASA Technical Reports Server (NTRS)

Niebur, D.; Germond, A.

1993-01-01

This report investigates the classification of power system states using an artificial neural network model, Kohonen's self-organizing feature map. The ultimate goal of this classification is to assess power system static security in real-time. Kohonen's self-organizing feature map is an unsupervised neural network which maps N-dimensional input vectors to an array of M neurons. After learning, the synaptic weight vectors exhibit a topological organization which represents the relationship between the vectors of the training set. This learning is unsupervised, which means that the number and size of the classes are not specified beforehand. In the application developed in this report, the input vectors used as the training set are generated by off-line load-flow simulations. The learning algorithm and the results of the organization are discussed.
Hyperspectral Imaging Using Flexible Endoscopy for Laryngeal Cancer Detection

PubMed Central

Regeling, Bianca; Thies, Boris; Gerstner, Andreas O. H.; Westermann, Stephan; Müller, Nina A.; Bendix, Jörg; Laffers, Wiebke

2016-01-01

Hyperspectral imaging (HSI) is increasingly gaining acceptance in the medical field. Up until now, HSI has been used in conjunction with rigid endoscopy to detect cancer in vivo. The logical next step is to pair HSI with flexible endoscopy, since it improves access to hard-to-reach areas. While the flexible endoscope’s fiber optic cables provide the advantage of flexibility, they also introduce an interfering honeycomb-like pattern onto images. Due to the substantial impact this pattern has on locating cancerous tissue, it must be removed before the HS data can be further processed. Thereby, the loss of information is to minimize avoiding the suppression of small-area variations of pixel values. We have developed a system that uses flexible endoscopy to record HS cubes of the larynx and designed a special filtering technique to remove the honeycomb-like pattern with minimal loss of information. We have confirmed its feasibility by comparing it to conventional filtering techniques using an objective metric and by applying unsupervised and supervised classifications to raw and pre-processed HS cubes. Compared to conventional techniques, our method successfully removes the honeycomb-like pattern and considerably improves classification performance, while preserving image details. PMID:27529255
Hyperspectral Imaging Using Flexible Endoscopy for Laryngeal Cancer Detection.

PubMed

Regeling, Bianca; Thies, Boris; Gerstner, Andreas O H; Westermann, Stephan; Müller, Nina A; Bendix, Jörg; Laffers, Wiebke

2016-08-13

Hyperspectral imaging (HSI) is increasingly gaining acceptance in the medical field. Up until now, HSI has been used in conjunction with rigid endoscopy to detect cancer in vivo. The logical next step is to pair HSI with flexible endoscopy, since it improves access to hard-to-reach areas. While the flexible endoscope's fiber optic cables provide the advantage of flexibility, they also introduce an interfering honeycomb-like pattern onto images. Due to the substantial impact this pattern has on locating cancerous tissue, it must be removed before the HS data can be further processed. Thereby, the loss of information is to minimize avoiding the suppression of small-area variations of pixel values. We have developed a system that uses flexible endoscopy to record HS cubes of the larynx and designed a special filtering technique to remove the honeycomb-like pattern with minimal loss of information. We have confirmed its feasibility by comparing it to conventional filtering techniques using an objective metric and by applying unsupervised and supervised classifications to raw and pre-processed HS cubes. Compared to conventional techniques, our method successfully removes the honeycomb-like pattern and considerably improves classification performance, while preserving image details.
Automated classifications of topography from DEMs by an unsupervised nested-means algorithm and a three-part geometric signature

NASA Astrophysics Data System (ADS)

Iwahashi, Junko; Pike, Richard J.

2007-05-01

An iterative procedure that implements the classification of continuous topography as a problem in digital image-processing automatically divides an area into categories of surface form; three taxonomic criteria-slope gradient, local convexity, and surface texture-are calculated from a square-grid digital elevation model (DEM). The sequence of programmed operations combines twofold-partitioned maps of the three variables converted to greyscale images, using the mean of each variable as the dividing threshold. To subdivide increasingly subtle topography, grid cells sloping at less than mean gradient of the input DEM are classified by designating mean values of successively lower-sloping subsets of the study area (nested means) as taxonomic thresholds, thereby increasing the number of output categories from the minimum 8 to 12 or 16. Program output is exemplified by 16 topographic types for the world at 1-km spatial resolution (SRTM30 data), the Japanese Islands at 270 m, and part of Hokkaido at 55 m. Because the procedure is unsupervised and reflects frequency distributions of the input variables rather than pre-set criteria, the resulting classes are undefined and must be calibrated empirically by subsequent analysis. Maps of the example classifications reflect physiographic regions, geological structure, and landform as well as slope materials and processes; fine-textured terrain categories tend to correlate with erosional topography or older surfaces, coarse-textured classes with areas of little dissection. In Japan the resulting classes approximate landform types mapped from airphoto analysis, while in the Americas they create map patterns resembling Hammond's terrain types or surface-form classes; SRTM30 output for the United States compares favorably with Fenneman's physical divisions. Experiments are suggested for further developing the method; the Arc/Info AML and the map of terrain classes for the world are available as online downloads.
Automated classifications of topography from DEMs by an unsupervised nested-means algorithm and a three-part geometric signature

USGS Publications Warehouse

Iwahashi, J.; Pike, R.J.

2007-01-01

An iterative procedure that implements the classification of continuous topography as a problem in digital image-processing automatically divides an area into categories of surface form; three taxonomic criteria-slope gradient, local convexity, and surface texture-are calculated from a square-grid digital elevation model (DEM). The sequence of programmed operations combines twofold-partitioned maps of the three variables converted to greyscale images, using the mean of each variable as the dividing threshold. To subdivide increasingly subtle topography, grid cells sloping at less than mean gradient of the input DEM are classified by designating mean values of successively lower-sloping subsets of the study area (nested means) as taxonomic thresholds, thereby increasing the number of output categories from the minimum 8 to 12 or 16. Program output is exemplified by 16 topographic types for the world at 1-km spatial resolution (SRTM30 data), the Japanese Islands at 270??m, and part of Hokkaido at 55??m. Because the procedure is unsupervised and reflects frequency distributions of the input variables rather than pre-set criteria, the resulting classes are undefined and must be calibrated empirically by subsequent analysis. Maps of the example classifications reflect physiographic regions, geological structure, and landform as well as slope materials and processes; fine-textured terrain categories tend to correlate with erosional topography or older surfaces, coarse-textured classes with areas of little dissection. In Japan the resulting classes approximate landform types mapped from airphoto analysis, while in the Americas they create map patterns resembling Hammond's terrain types or surface-form classes; SRTM30 output for the United States compares favorably with Fenneman's physical divisions. Experiments are suggested for further developing the method; the Arc/Info AML and the map of terrain classes for the world are available as online downloads. ?? 2006 Elsevier B.V. All rights reserved.
Learning from label proportions in brain-computer interfaces: Online unsupervised learning with guarantees.

PubMed

Hübner, David; Verhoeven, Thibault; Schmid, Konstantin; Müller, Klaus-Robert; Tangermann, Michael; Kindermans, Pieter-Jan

2017-01-01

Using traditional approaches, a brain-computer interface (BCI) requires the collection of calibration data for new subjects prior to online use. Calibration time can be reduced or eliminated e.g., by subject-to-subject transfer of a pre-trained classifier or unsupervised adaptive classification methods which learn from scratch and adapt over time. While such heuristics work well in practice, none of them can provide theoretical guarantees. Our objective is to modify an event-related potential (ERP) paradigm to work in unison with the machine learning decoder, and thus to achieve a reliable unsupervised calibrationless decoding with a guarantee to recover the true class means. We introduce learning from label proportions (LLP) to the BCI community as a new unsupervised, and easy-to-implement classification approach for ERP-based BCIs. The LLP estimates the mean target and non-target responses based on known proportions of these two classes in different groups of the data. We present a visual ERP speller to meet the requirements of LLP. For evaluation, we ran simulations on artificially created data sets and conducted an online BCI study with 13 subjects performing a copy-spelling task. Theoretical considerations show that LLP is guaranteed to minimize the loss function similar to a corresponding supervised classifier. LLP performed well in simulations and in the online application, where 84.5% of characters were spelled correctly on average without prior calibration. The continuously adapting LLP classifier is the first unsupervised decoder for ERP BCIs guaranteed to find the optimal decoder. This makes it an ideal solution to avoid tedious calibration sessions. Additionally, LLP works on complementary principles compared to existing unsupervised methods, opening the door for their further enhancement when combined with LLP.
Learning from label proportions in brain-computer interfaces: Online unsupervised learning with guarantees

PubMed Central

Verhoeven, Thibault; Schmid, Konstantin; Müller, Klaus-Robert; Tangermann, Michael; Kindermans, Pieter-Jan

2017-01-01

Objective Using traditional approaches, a brain-computer interface (BCI) requires the collection of calibration data for new subjects prior to online use. Calibration time can be reduced or eliminated e.g., by subject-to-subject transfer of a pre-trained classifier or unsupervised adaptive classification methods which learn from scratch and adapt over time. While such heuristics work well in practice, none of them can provide theoretical guarantees. Our objective is to modify an event-related potential (ERP) paradigm to work in unison with the machine learning decoder, and thus to achieve a reliable unsupervised calibrationless decoding with a guarantee to recover the true class means. Method We introduce learning from label proportions (LLP) to the BCI community as a new unsupervised, and easy-to-implement classification approach for ERP-based BCIs. The LLP estimates the mean target and non-target responses based on known proportions of these two classes in different groups of the data. We present a visual ERP speller to meet the requirements of LLP. For evaluation, we ran simulations on artificially created data sets and conducted an online BCI study with 13 subjects performing a copy-spelling task. Results Theoretical considerations show that LLP is guaranteed to minimize the loss function similar to a corresponding supervised classifier. LLP performed well in simulations and in the online application, where 84.5% of characters were spelled correctly on average without prior calibration. Significance The continuously adapting LLP classifier is the first unsupervised decoder for ERP BCIs guaranteed to find the optimal decoder. This makes it an ideal solution to avoid tedious calibration sessions. Additionally, LLP works on complementary principles compared to existing unsupervised methods, opening the door for their further enhancement when combined with LLP. PMID:28407016
Application of satellite data and LARS's data processing techniques to mapping vegetation of the Dismal Swamp. M.S. Thesis - Old Dominion Univ.

NASA Technical Reports Server (NTRS)

Messmore, J. A.

1976-01-01

The feasibility of using digital satellite imagery and automatic data processing techniques as a means of mapping swamp forest vegetation was considered, using multispectral scanner data acquired by the LANDSAT-1 satellite. The site for this investigation was the Dismal Swamp, a 210,000 acre swamp forest located south of Suffolk, Va. on the Virginia-North Carolina border. Two basic classification strategies were employed. The initial classification utilized unsupervised techniques which produced a map of the swamp indicating the distribution of thirteen forest spectral classes. These classes were later combined into three informational categories: Atlantic white cedar (Chamaecyparis thyoides), Loblolly pine (Pinus taeda), and deciduous forest. The subsequent classification employed supervised techniques which mapped Atlantic white cedar, Loblolly pine, deciduous forest, water and agriculture within the study site. A classification accuracy of 82.5% was produced by unsupervised techniques compared with 89% accuracy using supervised techniques.
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning.

PubMed

Shin, Hoo-Chang; Roth, Holger R; Gao, Mingchen; Lu, Le; Xu, Ziyue; Nogues, Isabella; Yao, Jianhua; Mollura, Daniel; Summers, Ronald M

2016-05-01

Remarkable progress has been made in image recognition, primarily due to the availability of large-scale annotated datasets and deep convolutional neural networks (CNNs). CNNs enable learning data-driven, highly representative, hierarchical image features from sufficient training data. However, obtaining datasets as comprehensively annotated as ImageNet in the medical imaging domain remains a challenge. There are currently three major techniques that successfully employ CNNs to medical image classification: training the CNN from scratch, using off-the-shelf pre-trained CNN features, and conducting unsupervised CNN pre-training with supervised fine-tuning. Another effective method is transfer learning, i.e., fine-tuning CNN models pre-trained from natural image dataset to medical image tasks. In this paper, we exploit three important, but previously understudied factors of employing deep convolutional neural networks to computer-aided detection problems. We first explore and evaluate different CNN architectures. The studied models contain 5 thousand to 160 million parameters, and vary in numbers of layers. We then evaluate the influence of dataset scale and spatial image context on performance. Finally, we examine when and why transfer learning from pre-trained ImageNet (via fine-tuning) can be useful. We study two specific computer-aided detection (CADe) problems, namely thoraco-abdominal lymph node (LN) detection and interstitial lung disease (ILD) classification. We achieve the state-of-the-art performance on the mediastinal LN detection, and report the first five-fold cross-validation classification results on predicting axial CT slices with ILD categories. Our extensive empirical evaluation, CNN model analysis and valuable insights can be extended to the design of high performance CAD systems for other medical imaging tasks.
Ensemble Semi-supervised Frame-work for Brain Magnetic Resonance Imaging Tissue Segmentation

PubMed Central

Azmi, Reza; Pishgoo, Boshra; Norozi, Narges; Yeganeh, Samira

2013-01-01

Brain magnetic resonance images (MRIs) tissue segmentation is one of the most important parts of the clinical diagnostic tools. Pixel classification methods have been frequently used in the image segmentation with two supervised and unsupervised approaches up to now. Supervised segmentation methods lead to high accuracy, but they need a large amount of labeled data, which is hard, expensive, and slow to obtain. Moreover, they cannot use unlabeled data to train classifiers. On the other hand, unsupervised segmentation methods have no prior knowledge and lead to low level of performance. However, semi-supervised learning which uses a few labeled data together with a large amount of unlabeled data causes higher accuracy with less trouble. In this paper, we propose an ensemble semi-supervised frame-work for segmenting of brain magnetic resonance imaging (MRI) tissues that it has been used results of several semi-supervised classifiers simultaneously. Selecting appropriate classifiers has a significant role in the performance of this frame-work. Hence, in this paper, we present two semi-supervised algorithms expectation filtering maximization and MCo_Training that are improved versions of semi-supervised methods expectation maximization and Co_Training and increase segmentation accuracy. Afterward, we use these improved classifiers together with graph-based semi-supervised classifier as components of the ensemble frame-work. Experimental results show that performance of segmentation in this approach is higher than both supervised methods and the individual semi-supervised classifiers. PMID:24098863

Spectrally based mapping of riverbed composition

USGS Publications Warehouse

Legleiter, Carl; Stegman, Tobin K.; Overstreet, Brandon T.

2016-01-01

Remote sensing methods provide an efficient means of characterizing fluvial systems. This study evaluated the potential to map riverbed composition based on in situ and/or remote measurements of reflectance. Field spectra and substrate photos from the Snake River, Wyoming, USA, were used to identify different sediment facies and degrees of algal development and to quantify their optical characteristics. We hypothesized that accounting for the effects of depth and water column attenuation to isolate the reflectance of the streambed would enhance distinctions among bottom types and facilitate substrate classification. A bottom reflectance retrieval algorithm adapted from coastal research yielded realistic spectra for the 450 to 700 nm range; but bottom reflectance-based substrate classifications, generated using a random forest technique, were no more accurate than classifications derived from above-water field spectra. Additional hypothesis testing indicated that a combination of reflectance magnitude (brightness) and indices of spectral shape provided the most accurate riverbed classifications. Convolving field spectra to the response functions of a multispectral satellite and a hyperspectral imaging system did not reduce classification accuracies, implying that high spectral resolution was not essential. Supervised classifications of algal density produced from hyperspectral data and an inferred bottom reflectance image were not highly accurate, but unsupervised classification of the bottom reflectance image revealed distinct spectrally based clusters, suggesting that such an image could provide additional river information. We attribute the failure of bottom reflectance retrieval to yield more reliable substrate maps to a latent correlation between depth and bottom type. Accounting for the effects of depth might have eliminated a key distinction among substrates and thus reduced discriminatory power. Although further, more systematic study across a broader range of fluvial environments is needed to substantiate our initial results, this case study suggests that bed composition in shallow, clear-flowing rivers potentially could be mapped remotely.
EL68D Wasteway Watershed Land-Cover Generation

USGS Publications Warehouse

Ruhl, Sheila; Usery, E. Lynn; Finn, Michael P.

2007-01-01

Classification of land cover from Landsat Enhanced Thematic Mapper Plus (ETM+) for the EL68D Wasteway Watershed in the State of Washington is documented. The procedures for classification include use of two ETM+ scenes in a simultaneous unsupervised classification process supported by extensive field data collection using Global Positioning System receivers and digital photos. The procedure resulted in a detailed classification at the individual crop species level.
Linear mixing model applied to AVHRR LAC data

NASA Technical Reports Server (NTRS)

Holben, Brent N.; Shimabukuro, Yosio E.

1993-01-01

A linear mixing model was applied to coarse spatial resolution data from the NOAA Advanced Very High Resolution Radiometer. The reflective component of the 3.55 - 3.93 microns channel was extracted and used with the two reflective channels 0.58 - 0.68 microns and 0.725 - 1.1 microns to run a Constraine Least Squares model to generate vegetation, soil, and shade fraction images for an area in the Western region of Brazil. The Landsat Thematic Mapper data covering the Emas National park region was used for estimating the spectral response of the mixture components and for evaluating the mixing model results. The fraction images were compared with an unsupervised classification derived from Landsat TM data acquired on the same day. The relationship between the fraction images and normalized difference vegetation index images show the potential of the unmixing techniques when using coarse resolution data for global studies.
Unsupervised Deep Hashing With Pseudo Labels for Scalable Image Retrieval.

PubMed

Zhang, Haofeng; Liu, Li; Long, Yang; Shao, Ling

2018-04-01

In order to achieve efficient similarity searching, hash functions are designed to encode images into low-dimensional binary codes with the constraint that similar features will have a short distance in the projected Hamming space. Recently, deep learning-based methods have become more popular, and outperform traditional non-deep methods. However, without label information, most state-of-the-art unsupervised deep hashing (DH) algorithms suffer from severe performance degradation for unsupervised scenarios. One of the main reasons is that the ad-hoc encoding process cannot properly capture the visual feature distribution. In this paper, we propose a novel unsupervised framework that has two main contributions: 1) we convert the unsupervised DH model into supervised by discovering pseudo labels; 2) the framework unifies likelihood maximization, mutual information maximization, and quantization error minimization so that the pseudo labels can maximumly preserve the distribution of visual features. Extensive experiments on three popular data sets demonstrate the advantages of the proposed method, which leads to significant performance improvement over the state-of-the-art unsupervised hashing algorithms.
A study of earthquake-induced building detection by object oriented classification approach

NASA Astrophysics Data System (ADS)

Sabuncu, Asli; Damla Uca Avci, Zehra; Sunar, Filiz

2017-04-01

Among the natural hazards, earthquakes are the most destructive disasters and cause huge loss of lives, heavily infrastructure damages and great financial losses every year all around the world. According to the statistics about the earthquakes, more than a million earthquakes occur which is equal to two earthquakes per minute in the world. Natural disasters have brought more than 780.000 deaths approximately % 60 of all mortality is due to the earthquakes after 2001. A great earthquake took place at 38.75 N 43.36 E in the eastern part of Turkey in Van Province on On October 23th, 2011. 604 people died and about 4000 buildings seriously damaged and collapsed after this earthquake. In recent years, the use of object oriented classification approach based on different object features, such as spectral, textural, shape and spatial information, has gained importance and became widespread for the classification of high-resolution satellite images and orthophotos. The motivation of this study is to detect the collapsed buildings and debris areas after the earthquake by using very high-resolution satellite images and orthophotos with the object oriented classification and also see how well remote sensing technology was carried out in determining the collapsed buildings. In this study, two different land surfaces were selected as homogenous and heterogeneous case study areas. In the first step of application, multi-resolution segmentation was applied and optimum parameters were selected to obtain the objects in each area after testing different color/shape and compactness/smoothness values. In the next step, two different classification approaches, namely "supervised" and "unsupervised" approaches were applied and their classification performances were compared. Object-based Image Analysis (OBIA) was performed using e-Cognition software.
Automatic Galaxy Classification via Machine Learning Techniques: Parallelized Rotation/Flipping INvariant Kohonen Maps (PINK)

NASA Astrophysics Data System (ADS)

Polsterer, K. L.; Gieseke, F.; Igel, C.

2015-09-01

In the last decades more and more all-sky surveys created an enormous amount of data which is publicly available on the Internet. Crowd-sourcing projects such as Galaxy-Zoo and Radio-Galaxy-Zoo used encouraged users from all over the world to manually conduct various classification tasks. The combination of the pattern-recognition capabilities of thousands of volunteers enabled scientists to finish the data analysis within acceptable time. For up-coming surveys with billions of sources, however, this approach is not feasible anymore. In this work, we present an unsupervised method that can automatically process large amounts of galaxy data and which generates a set of prototypes. This resulting model can be used to both visualize the given galaxy data as well as to classify so far unseen images.
Unsupervised Feature Learning for Heart Sounds Classification Using Autoencoder

NASA Astrophysics Data System (ADS)

Hu, Wei; Lv, Jiancheng; Liu, Dongbo; Chen, Yao

2018-04-01

Cardiovascular disease seriously threatens the health of many people. It is usually diagnosed during cardiac auscultation, which is a fast and efficient method of cardiovascular disease diagnosis. In recent years, deep learning approach using unsupervised learning has made significant breakthroughs in many fields. However, to our knowledge, deep learning has not yet been used for heart sound classification. In this paper, we first use the average Shannon energy to extract the envelope of the heart sounds, then find the highest point of S1 to extract the cardiac cycle. We convert the time-domain signals of the cardiac cycle into spectrograms and apply principal component analysis whitening to reduce the dimensionality of the spectrogram. Finally, we apply a two-layer autoencoder to extract the features of the spectrogram. The experimental results demonstrate that the features from the autoencoder are suitable for heart sound classification.
Modeling EEG Waveforms with Semi-Supervised Deep Belief Nets: Fast Classification and Anomaly Measurement

PubMed Central

Wulsin, D. F.; Gupta, J. R.; Mani, R.; Blanco, J. A.; Litt, B.

2011-01-01

Clinical electroencephalography (EEG) records vast amounts of human complex data yet is still reviewed primarily by human readers. Deep Belief Nets (DBNs) are a relatively new type of multi-layer neural network commonly tested on two-dimensional image data, but are rarely applied to times-series data such as EEG. We apply DBNs in a semi-supervised paradigm to model EEG waveforms for classification and anomaly detection. DBN performance was comparable to standard classifiers on our EEG dataset, and classification time was found to be 1.7 to 103.7 times faster than the other high-performing classifiers. We demonstrate how the unsupervised step of DBN learning produces an autoencoder that can naturally be used in anomaly measurement. We compare the use of raw, unprocessed data—a rarity in automated physiological waveform analysis—to hand-chosen features and find that raw data produces comparable classification and better anomaly measurement performance. These results indicate that DBNs and raw data inputs may be more effective for online automated EEG waveform recognition than other common techniques. PMID:21525569
Towards a robust framework for catchment classification

NASA Astrophysics Data System (ADS)

Deshmukh, A.; Samal, A.; Singh, R.

2017-12-01

Classification of catchments based on various measures of similarity has emerged as an important technique to understand regional scale hydrologic behavior. Classification of catchment characteristics and/or streamflow response has been used reveal which characteristics are more likely to explain the observed variability of hydrologic response. However, numerous algorithms for supervised or unsupervised classification are available, making it hard to identify the algorithm most suitable for the dataset at hand. Consequently, existing catchment classification studies vary significantly in the classification algorithms employed with no previous attempt at understanding the degree of uncertainty in classification due to this algorithmic choice. This hinders the generalizability of interpretations related to hydrologic behavior. Our goal is to develop a protocol that can be followed while classifying hydrologic datasets. We focus on a classification framework for unsupervised classification and provide a step-by-step classification procedure. The steps include testing the clusterabiltiy of original dataset prior to classification, feature selection, validation of clustered data, and quantification of similarity of two clusterings. We test several commonly available methods within this framework to understand the level of similarity of classification results across algorithms. We apply the proposed framework on recently developed datasets for India to analyze to what extent catchment properties can explain observed catchment response. Our testing dataset includes watershed characteristics for over 200 watersheds which comprise of both natural (physio-climatic) characteristics and socio-economic characteristics. This framework allows us to understand the controls on observed hydrologic variability across India.
A comparison of neural network and fuzzy clustering techniques in segmenting magnetic resonance images of the brain.

PubMed

Hall, L O; Bensaid, A M; Clarke, L P; Velthuizen, R P; Silbiger, M S; Bezdek, J C

1992-01-01

Magnetic resonance (MR) brain section images are segmented and then synthetically colored to give visual representations of the original data with three approaches: the literal and approximate fuzzy c-means unsupervised clustering algorithms, and a supervised computational neural network. Initial clinical results are presented on normal volunteers and selected patients with brain tumors surrounded by edema. Supervised and unsupervised segmentation techniques provide broadly similar results. Unsupervised fuzzy algorithms were visually observed to show better segmentation when compared with raw image data for volunteer studies. For a more complex segmentation problem with tumor/edema or cerebrospinal fluid boundary, where the tissues have similar MR relaxation behavior, inconsistency in rating among experts was observed, with fuzz-c-means approaches being slightly preferred over feedforward cascade correlation results. Various facets of both approaches, such as supervised versus unsupervised learning, time complexity, and utility for the diagnostic process, are compared.
Predicting category intuitiveness with the rational model, the simplicity model, and the generalized context model.

PubMed

Pothos, Emmanuel M; Bailey, Todd M

2009-07-01

Naïve observers typically perceive some groupings for a set of stimuli as more intuitive than others. The problem of predicting category intuitiveness has been historically considered the remit of models of unsupervised categorization. In contrast, this article develops a measure of category intuitiveness from one of the most widely supported models of supervised categorization, the generalized context model (GCM). Considering different category assignments for a set of instances, the authors asked how well the GCM can predict the classification of each instance on the basis of all the other instances. The category assignment that results in the smallest prediction error is interpreted as the most intuitive for the GCM-the authors refer to this way of applying the GCM as "unsupervised GCM." The authors systematically compared predictions of category intuitiveness from the unsupervised GCM and two models of unsupervised categorization: the simplicity model and the rational model. The unsupervised GCM compared favorably with the simplicity model and the rational model. This success of the unsupervised GCM illustrates that the distinction between supervised and unsupervised categorization may need to be reconsidered. However, no model emerged as clearly superior, indicating that there is more work to be done in understanding and modeling category intuitiveness.
Unsupervised detection of salt marsh platforms: a topographic method

NASA Astrophysics Data System (ADS)

Goodwin, Guillaume C. H.; Mudd, Simon M.; Clubb, Fiona J.

2018-03-01

Salt marshes filter pollutants, protect coastlines against storm surges, and sequester carbon, yet are under threat from sea level rise and anthropogenic modification. The sustained existence of the salt marsh ecosystem depends on the topographic evolution of marsh platforms. Quantifying marsh platform topography is vital for improving the management of these valuable landscapes. The determination of platform boundaries currently relies on supervised classification methods requiring near-infrared data to detect vegetation, or demands labour-intensive field surveys and digitisation. We propose a novel, unsupervised method to reproducibly isolate salt marsh scarps and platforms from a digital elevation model (DEM), referred to as Topographic Identification of Platforms (TIP). Field observations and numerical models show that salt marshes mature into subhorizontal platforms delineated by subvertical scarps. Based on this premise, we identify scarps as lines of local maxima on a slope raster, then fill landmasses from the scarps upward, thus isolating mature marsh platforms. We test the TIP method using lidar-derived DEMs from six salt marshes in England with varying tidal ranges and geometries, for which topographic platforms were manually isolated from tidal flats. Agreement between manual and unsupervised classification exceeds 94 % for DEM resolutions of 1 m, with all but one site maintaining an accuracy superior to 90 % for resolutions up to 3 m. For resolutions of 1 m, platforms detected with the TIP method are comparable in surface area to digitised platforms and have similar elevation distributions. We also find that our method allows for the accurate detection of local block failures as small as 3 times the DEM resolution. Detailed inspection reveals that although tidal creeks were digitised as part of the marsh platform, unsupervised classification categorises them as part of the tidal flat, causing an increase in false negatives and overall platform perimeter. This suggests our method may benefit from combination with existing creek detection algorithms. Fallen blocks and high tidal flat portions, associated with potential pioneer zones, can also lead to differences between our method and supervised mapping. Although pioneer zones prove difficult to classify using a topographic method, we suggest that these transition areas should be considered when analysing erosion and accretion processes, particularly in the case of incipient marsh platforms. Ultimately, we have shown that unsupervised classification of marsh platforms from high-resolution topography is possible and sufficient to monitor and analyse topographic evolution.
Automated connectionist-geostatistical classification as an approach to identify sea ice and land ice types, properties and provinces

NASA Astrophysics Data System (ADS)

Goetz-Weiss, L. R.; Herzfeld, U. C.; Trantow, T.; Hunke, E. C.; Maslanik, J. A.; Crocker, R. I.

2016-12-01

An important problem in model-data comparison is the identification of parameters that can be extracted from observational data as well as used in numerical models, which are typically based on idealized physical processes. Here, we present a suite of approaches to characterization and classification of sea ice and land ice types, properties and provinces based on several types of remote-sensing data. Applications will be given to not only illustrate the approach, but employ it in model evaluation and understanding of physical processes. (1) In a geostatistical characterization, spatial sea-ice properties in the Chukchi and Beaufort Sea and in Elsoon Lagoon are derived from analysis of RADARSAT and ERS-2 SAR data. (2) The analysis is taken further by utilizing multi-parameter feature vectors as inputs for unsupervised and supervised statistical classification, which facilitates classification of different sea-ice types. (3) Characteristic sea-ice parameters, as resultant from the classification, can then be applied in model evaluation, as demonstrated for the ridging scheme of the Los Alamos sea ice model, CICE, using high-resolution altimeter and image data collected from unmanned aircraft over Fram Strait during the Characterization of Arctic Sea Ice Experiment (CASIE). The characteristic parameters chosen in this application are directly related to deformation processes, which also underly the ridging scheme. (4) The method that is capable of the most complex classification tasks is the connectionist-geostatistical classification method. This approach has been developed to identify currently up to 18 different crevasse types in order to map progression of the surge through the complex Bering-Bagley Glacier System, Alaska, in 2011-2014. The analysis utilizes airborne altimeter data and video image data and satellite image data. Results of the crevasse classification are compare to fracture modeling and found to match.
Model-based approach to the detection and classification of mines in sidescan sonar.

PubMed

Reed, Scott; Petillot, Yvan; Bell, Judith

2004-01-10

This paper presents a model-based approach to mine detection and classification by use of sidescan sonar. Advances in autonomous underwater vehicle technology have increased the interest in automatic target recognition systems in an effort to automate a process that is currently carried out by a human operator. Current automated systems generally require training and thus produce poor results when the test data set is different from the training set. This has led to research into unsupervised systems, which are able to cope with the large variability in conditions and terrains seen in sidescan imagery. The system presented in this paper first detects possible minelike objects using a Markov random field model, which operates well on noisy images, such as sidescan, and allows a priori information to be included through the use of priors. The highlight and shadow regions of the object are then extracted with a cooperating statistical snake, which assumes these regions are statistically separate from the background. Finally, a classification decision is made using Dempster-Shafer theory, where the extracted features are compared with synthetic realizations generated with a sidescan sonar simulator model. Results for the entire process are shown on real sidescan sonar data. Similarities between the sidescan sonar and synthetic aperture radar (SAR) imaging processes ensure that the approach outlined here could be made applied to SAR image analysis.
Texture analysis based on the Hermite transform for image classification and segmentation

NASA Astrophysics Data System (ADS)

Estudillo-Romero, Alfonso; Escalante-Ramirez, Boris; Savage-Carmona, Jesus

2012-06-01

Texture analysis has become an important task in image processing because it is used as a preprocessing stage in different research areas including medical image analysis, industrial inspection, segmentation of remote sensed imaginary, multimedia indexing and retrieval. In order to extract visual texture features a texture image analysis technique is presented based on the Hermite transform. Psychovisual evidence suggests that the Gaussian derivatives fit the receptive field profiles of mammalian visual systems. The Hermite transform describes locally basic texture features in terms of Gaussian derivatives. Multiresolution combined with several analysis orders provides detection of patterns that characterizes every texture class. The analysis of the local maximum energy direction and steering of the transformation coefficients increase the method robustness against the texture orientation. This method presents an advantage over classical filter bank design because in the latter a fixed number of orientations for the analysis has to be selected. During the training stage, a subset of the Hermite analysis filters is chosen in order to improve the inter-class separability, reduce dimensionality of the feature vectors and computational cost during the classification stage. We exhaustively evaluated the correct classification rate of real randomly selected training and testing texture subsets using several kinds of common used texture features. A comparison between different distance measurements is also presented. Results of the unsupervised real texture segmentation using this approach and comparison with previous approaches showed the benefits of our proposal.
Video mining using combinations of unsupervised and supervised learning techniques

NASA Astrophysics Data System (ADS)

Divakaran, Ajay; Miyahara, Koji; Peker, Kadir A.; Radhakrishnan, Regunathan; Xiong, Ziyou

2003-12-01

We discuss the meaning and significance of the video mining problem, and present our work on some aspects of video mining. A simple definition of video mining is unsupervised discovery of patterns in audio-visual content. Such purely unsupervised discovery is readily applicable to video surveillance as well as to consumer video browsing applications. We interpret video mining as content-adaptive or "blind" content processing, in which the first stage is content characterization and the second stage is event discovery based on the characterization obtained in stage 1. We discuss the target applications and find that using a purely unsupervised approach are too computationally complex to be implemented on our product platform. We then describe various combinations of unsupervised and supervised learning techniques that help discover patterns that are useful to the end-user of the application. We target consumer video browsing applications such as commercial message detection, sports highlights extraction etc. We employ both audio and video features. We find that supervised audio classification combined with unsupervised unusual event discovery enables accurate supervised detection of desired events. Our techniques are computationally simple and robust to common variations in production styles etc.
Change detection and change monitoring of natural and man-made features in multispectral and hyperspectral satellite imagery

DOE Office of Scientific and Technical Information (OSTI.GOV)

Moody, Daniela Irina

An approach for land cover classification, seasonal and yearly change detection and monitoring, and identification of changes in man-made features may use a clustering of sparse approximations (CoSA) on sparse representations in learned dictionaries. A Hebbian learning rule may be used to build multispectral or hyperspectral, multiresolution dictionaries that are adapted to regional satellite image data. Sparse image representations of pixel patches over the learned dictionaries may be used to perform unsupervised k-means clustering into land cover categories. The clustering process behaves as a classifier in detecting real variability. This approach may combine spectral and spatial textural characteristics to detectmore » geologic, vegetative, hydrologic, and man-made features, as well as changes in these features over time.« less
Advances in optical information processing IV; Proceedings of the Meeting, Orlando, FL, Apr. 18-20, 1990

NASA Astrophysics Data System (ADS)

Pape, Dennis R.

1990-09-01

The present conference discusses topics in optical image processing, optical signal processing, acoustooptic spectrum analyzer systems and components, and optical computing. Attention is given to tradeoffs in nonlinearly recorded matched filters, miniature spatial light modulators, detection and classification using higher-order statistics of optical matched filters, rapid traversal of an image data base using binary synthetic discriminant filters, wideband signal processing for emitter location, an acoustooptic processor for autonomous SAR guidance, and sampling of Fresnel transforms. Also discussed are an acoustooptic RF signal-acquisition system, scanning acoustooptic spectrum analyzers, the effects of aberrations on acoustooptic systems, fast optical digital arithmetic processors, information utilization in analog and digital processing, optical processors for smart structures, and a self-organizing neural network for unsupervised learning.
Automatic segmentation of brain MRI in high-dimensional local and non-local feature space based on sparse representation.

PubMed

Khalilzadeh, Mohammad Mahdi; Fatemizadeh, Emad; Behnam, Hamid

2013-06-01

Automatic extraction of the varying regions of magnetic resonance images is required as a prior step in a diagnostic intelligent system. The sparsest representation and high-dimensional feature are provided based on learned dictionary. The classification is done by employing the technique that computes the reconstruction error locally and non-locally of each pixel. The acquired results from the real and simulated images are superior to the best MRI segmentation method with regard to the stability advantages. In addition, it is segmented exactly through a formula taken from the distance and sparse factors. Also, it is done automatically taking sparse factor in unsupervised clustering methods whose results have been improved. Copyright © 2013 Elsevier Inc. All rights reserved.
Topic Model for Graph Mining.

PubMed

Xuan, Junyu; Lu, Jie; Zhang, Guangquan; Luo, Xiangfeng

2015-12-01

Graph mining has been a popular research area because of its numerous application scenarios. Many unstructured and structured data can be represented as graphs, such as, documents, chemical molecular structures, and images. However, an issue in relation to current research on graphs is that they cannot adequately discover the topics hidden in graph-structured data which can be beneficial for both the unsupervised learning and supervised learning of the graphs. Although topic models have proved to be very successful in discovering latent topics, the standard topic models cannot be directly applied to graph-structured data due to the "bag-of-word" assumption. In this paper, an innovative graph topic model (GTM) is proposed to address this issue, which uses Bernoulli distributions to model the edges between nodes in a graph. It can, therefore, make the edges in a graph contribute to latent topic discovery and further improve the accuracy of the supervised and unsupervised learning of graphs. The experimental results on two different types of graph datasets show that the proposed GTM outperforms the latent Dirichlet allocation on classification by using the unveiled topics of these two models to represent graphs.

The DSFPN, a new neural network for optical character recognition.

PubMed

Morns, L P; Dlay, S S

1999-01-01

A new type of neural network for recognition tasks is presented in this paper. The network, called the dynamic supervised forward-propagation network (DSFPN), is based on the forward only version of the counterpropagation network (CPN). The DSFPN, trains using a supervised algorithm and can grow dynamically during training, allowing subclasses in the training data to be learnt in an unsupervised manner. It is shown to train in times comparable to the CPN while giving better classification accuracies than the popular backpropagation network. Both Fourier descriptors and wavelet descriptors are used for image preprocessing and the wavelets are proven to give a far better performance.
Remote sensing of environmental impact of land use activities

NASA Technical Reports Server (NTRS)

Paul, C. K.

1977-01-01

The capability to monitor land cover, associated in the past with aerial film cameras and radar systems, was discussed in regard to aircraft and spacecraft multispectral scanning sensors. A proposed thematic mapper with greater spectral and spatial resolutions for the fourth LANDSAT is expected to usher in new environmental monitoring capability. In addition, continuing improvements in image classification by supervised and unsupervised computer techniques are being operationally verified for discriminating environmental impacts of human activities on the land. The benefits of employing remote sensing for this discrimination was shown to far outweigh the incremental costs of converting to an aircraft-satellite multistage system.
V-Sipal - a Virtual Laboratory for Satellite Image Processing and Analysis

NASA Astrophysics Data System (ADS)

Buddhiraju, K. M.; Eeti, L.; Tiwari, K. K.

2011-09-01

In this paper a virtual laboratory for the Satellite Image Processing and Analysis (v-SIPAL) being developed at the Indian Institute of Technology Bombay is described. v-SIPAL comprises a set of experiments that are normally carried out by students learning digital processing and analysis of satellite images using commercial software. Currently, the experiments that are available on the server include Image Viewer, Image Contrast Enhancement, Image Smoothing, Edge Enhancement, Principal Component Transform, Texture Analysis by Co-occurrence Matrix method, Image Indices, Color Coordinate Transforms, Fourier Analysis, Mathematical Morphology, Unsupervised Image Classification, Supervised Image Classification and Accuracy Assessment. The virtual laboratory includes a theory module for each option of every experiment, a description of the procedure to perform each experiment, the menu to choose and perform the experiment, a module on interpretation of results when performed with a given image and pre-specified options, bibliography, links to useful internet resources and user-feedback. The user can upload his/her own images for performing the experiments and can also reuse outputs of one experiment in another experiment where applicable. Some of the other experiments currently under development include georeferencing of images, data fusion, feature evaluation by divergence andJ-M distance, image compression, wavelet image analysis and change detection. Additions to the theory module include self-assessment quizzes, audio-video clips on selected concepts, and a discussion of elements of visual image interpretation. V-SIPAL is at the satge of internal evaluation within IIT Bombay and will soon be open to selected educational institutions in India for evaluation.
Classification of neocortical interneurons using affinity propagation.

PubMed

Santana, Roberto; McGarry, Laura M; Bielza, Concha; Larrañaga, Pedro; Yuste, Rafael

2013-01-01

In spite of over a century of research on cortical circuits, it is still unknown how many classes of cortical neurons exist. In fact, neuronal classification is a difficult problem because it is unclear how to designate a neuronal cell class and what are the best characteristics to define them. Recently, unsupervised classifications using cluster analysis based on morphological, physiological, or molecular characteristics, have provided quantitative and unbiased identification of distinct neuronal subtypes, when applied to selected datasets. However, better and more robust classification methods are needed for increasingly complex and larger datasets. Here, we explored the use of affinity propagation, a recently developed unsupervised classification algorithm imported from machine learning, which gives a representative example or exemplar for each cluster. As a case study, we applied affinity propagation to a test dataset of 337 interneurons belonging to four subtypes, previously identified based on morphological and physiological characteristics. We found that affinity propagation correctly classified most of the neurons in a blind, non-supervised manner. Affinity propagation outperformed Ward's method, a current standard clustering approach, in classifying the neurons into 4 subtypes. Affinity propagation could therefore be used in future studies to validly classify neurons, as a first step to help reverse engineer neural circuits.
Classification of damage in structural systems using time series analysis and supervised and unsupervised pattern recognition techniques

NASA Astrophysics Data System (ADS)

Omenzetter, Piotr; de Lautour, Oliver R.

2010-04-01

Developed for studying long, periodic records of various measured quantities, time series analysis methods are inherently suited and offer interesting possibilities for Structural Health Monitoring (SHM) applications. However, their use in SHM can still be regarded as an emerging application and deserves more studies. In this research, Autoregressive (AR) models were used to fit experimental acceleration time histories from two experimental structural systems, a 3- storey bookshelf-type laboratory structure and the ASCE Phase II SHM Benchmark Structure, in healthy and several damaged states. The coefficients of the AR models were chosen as damage sensitive features. Preliminary visual inspection of the large, multidimensional sets of AR coefficients to check the presence of clusters corresponding to different damage severities was achieved using Sammon mapping - an efficient nonlinear data compression technique. Systematic classification of damage into states based on the analysis of the AR coefficients was achieved using two supervised classification techniques: Nearest Neighbor Classification (NNC) and Learning Vector Quantization (LVQ), and one unsupervised technique: Self-organizing Maps (SOM). This paper discusses the performance of AR coefficients as damage sensitive features and compares the efficiency of the three classification techniques using experimental data.
Single-particle cryo-EM using alignment by classification (ABC): the structure of Lumbricus terrestris haemoglobin

PubMed Central

Seer-Linnemayr, Charlotte; Ravelli, Raimond B. G.; Matadeen, Rishi; De Carlo, Sacha; Alewijnse, Bart; Portugal, Rodrigo V.; Pannu, Navraj S.; Schatz, Michael; van Heel, Marin

2017-01-01

Single-particle cryogenic electron microscopy (cryo-EM) can now yield near-atomic resolution structures of biological complexes. However, the reference-based alignment algorithms commonly used in cryo-EM suffer from reference bias, limiting their applicability (also known as the ‘Einstein from random noise’ problem). Low-dose cryo-EM therefore requires robust and objective approaches to reveal the structural information contained in the extremely noisy data, especially when dealing with small structures. A reference-free pipeline is presented for obtaining near-atomic resolution three-dimensional reconstructions from heterogeneous (‘four-dimensional’) cryo-EM data sets. The methodologies integrated in this pipeline include a posteriori camera correction, movie-based full-data-set contrast transfer function determination, movie-alignment algorithms, (Fourier-space) multivariate statistical data compression and unsupervised classification, ‘random-startup’ three-dimensional reconstructions, four-dimensional structural refinements and Fourier shell correlation criteria for evaluating anisotropic resolution. The procedures exclusively use information emerging from the data set itself, without external ‘starting models’. Euler-angle assignments are performed by angular reconstitution rather than by the inherently slower projection-matching approaches. The comprehensive ‘ABC-4D’ pipeline is based on the two-dimensional reference-free ‘alignment by classification’ (ABC) approach, where similar images in similar orientations are grouped by unsupervised classification. Some fundamental differences between X-ray crystallography versus single-particle cryo-EM data collection and data processing are discussed. The structure of the giant haemoglobin from Lumbricus terrestris at a global resolution of ∼3.8 Å is presented as an example of the use of the ABC-4D procedure. PMID:28989723
Quantification and Segmentation of Brain Tissues from MR Images: A Probabilistic Neural Network Approach

PubMed Central

Wang, Yue; Adalý, Tülay; Kung, Sun-Yuan; Szabo, Zsolt

2007-01-01

This paper presents a probabilistic neural network based technique for unsupervised quantification and segmentation of brain tissues from magnetic resonance images. It is shown that this problem can be solved by distribution learning and relaxation labeling, resulting in an efficient method that may be particularly useful in quantifying and segmenting abnormal brain tissues where the number of tissue types is unknown and the distributions of tissue types heavily overlap. The new technique uses suitable statistical models for both the pixel and context images and formulates the problem in terms of model-histogram fitting and global consistency labeling. The quantification is achieved by probabilistic self-organizing mixtures and the segmentation by a probabilistic constraint relaxation network. The experimental results show the efficient and robust performance of the new algorithm and that it outperforms the conventional classification based approaches. PMID:18172510
Regional Climate Modeling over the Marmara Region, Turkey, with Improved Land Cover Data

NASA Astrophysics Data System (ADS)

Sertel, E.; Robock, A.

2007-12-01

Land surface controls the partitioning of available energy at the surface between sensible and latent heat,and controls partitioning of available water between evaporation and runoff. Current land cover data available within the regional climate models such as Regional Atmospheric Modeling System (RAMS), the Fifth-Generation NCAR/Penn State Mesoscale Model (MM5) and Weather Research and Forecasting (WRF) was obtained from 1- km Advanced Very High Resolution Radiometer satellite images spanning April 1992 through March 1993 with an unsupervised classification technique. These data are not up-to-date and are not accurate for all regions and some land cover types such as urban areas. Here we introduce new, up-to-date and accurate land cover data for the Marmara Region, Turkey derived from Landsat Enhanced Thematic Mapper images into the WRF regional climate model. We used several image processing techniques to create accurate land cover data from Landsat images obtained between 2001 and 2005. First, all images were atmospherically and radiometrically corrected to minimize contamination effects of atmospheric particles and systematic errors. Then, geometric correction was performed for each image to eliminate geometric distortions and define images in a common coordinate system. Finally, unsupervised and supervised classification techniques were utilized to form the most accurate land cover data yet for the study area. Accuracy assessments of the classifications were performed using error matrix and kappa statistics to find the best classification results. Maximum likelihood classification method gave the most accurate results over the study area. We compared the new land cover data with the default WRF land cover data. WRF land cover data cannot represent urban areas in the cities of Istanbul, Izmit, and Bursa. As an example, both original satellite images and new land cover data showed the expansion of urban areas into the Istanbul metropolitan area, but in the WRF land cover data only a limited area along the Bosporus is shown as urban. In addition, the new land cover data indicate that the northern part of Istanbul is covered by evergreen and deciduous forest (verified by ground truth data), but the WRF data indicate that most of this region is croplands. In the northern part of the Marmara Region, there is bare ground as a result of open mining activities and this class can be identified in our land cover data, whereas the WRF data indicated this region as woodland. We then used this new data set to conduct WRF simulations for one main and two nested domains, where the inner-most domain represents the Marmara Region with 3 km horizontal resolution. The vertical domain of both main and nested domains extends over 28 vertical levels. Initial and boundary conditions were obtained from National Centers for Environmental Prediction-Department of Energy Reanalysis II and the Noah model was selected as the land surface model. Two model simulations were conducted; one with available land cover data and one with the newly created land cover data. Using detailed meteorological station data within the study area, we find that the simulation with the new land cover data set produces better temperature and precipitation simulations for the region, showing the value of accurate land cover data and that changing land cover data can be an important influence on local climate change.
Spectroscopic classification of icy satellites of Saturn II: Identification of terrain units on Rhea

NASA Astrophysics Data System (ADS)

Scipioni, F.; Tosi, F.; Stephan, K.; Filacchione, G.; Ciarniello, M.; Capaccioni, F.; Cerroni, P.

2014-05-01

Rhea is the second largest icy satellites of Saturn and it is mainly composed of water ice. Its surface is characterized by a leading hemisphere slightly brighter than the trailing side. The main goal of this work is to identify homogeneous compositional units on Rhea by applying the Spectral Angle Mapper (SAM) classification technique to Rhea’s hyperspectral images acquired by the Visual and Infrared Mapping Spectrometer (VIMS) onboard the Cassini Orbiter in the infrared range (0.88-5.12 μm). The first step of the classification is dedicated to the identification of Rhea’s spectral endmembers by applying the k-means unsupervised clustering technique to four hyperspectral images representative of a limited portion of the surface, imaged at relatively high spatial resolution. We then identified eight spectral endmembers, corresponding to as many terrain units, which mostly distinguish for water ice abundance and ice grain size. In the second step, endmembers are used as reference spectra in SAM classification method to achieve a comprehensive classification of the entire surface. From our analysis of the infrared spectra returned by VIMS, it clearly emerges that Rhea’ surface units shows differences in terms of water ice bands depths, average ice grain size, and concentration of contaminants, particularly CO2 and hydrocarbons. The spectral units that classify optically dark terrains are those showing suppressed water ice bands, a finer ice grain size and a higher concentration of carbon dioxide. Conversely, spectral units labeling brighter regions have deeper water ice absorption bands, higher albedo and a smaller concentration of contaminants. All these variations reflect surface’s morphological and geological structures. Finally, we performed a comparison between Rhea and Dione, to highlight different magnitudes of space weathering effects in the icy satellites as a function of the distance from Saturn.
Coral reef monitoring by the compact airborne spectrographic imager (CASI)

NASA Astrophysics Data System (ADS)

Miyazaki, Tadakuni; Tokumura, Kimiaki; Sugita, Mikio

1997-12-01

The casi has the spatial resolution of about 3 m X 3 m at the ground level and its spectral resolution is about 6 nm. The wavelength range for the measurement is from 430 to 870 nm and the number of the bands is 72 bands. An airplane carrying casi flew over Kuroshima Island, Okinawa and acquired image data of Kuroshima Island and the surrounding sea area. The flight courses were 6 courses at the altitude of 9,000 feet and 2 courses of 6,000 feet. At the same time, spectral measurements of the sea surface and several coral reefs underwater were carried out at an area of coral reefs off Kuroshima Island. The supervised and unsupervised classification were applied to the casi imageries to extract and classify the area of coral reefs off Kuroshima Island. The produced classification maps of the coral reefs were compared with the ground truth map of coral reefs made by the professional divers to evaluate the results. The results showed significant similarity of the distribution pattern of corral reefs.
Digital Image Sensor-Based Assessment of the Status of Oat (Avena sativa L.) Crops after Frost Damage

PubMed Central

Macedo-Cruz, Antonia; Pajares, Gonzalo; Santos, Matilde; Villegas-Romero, Isidro

2011-01-01

The aim of this paper is to classify the land covered with oat crops, and the quantification of frost damage on oats, while plants are still in the flowering stage. The images are taken by a digital colour camera CCD-based sensor. Unsupervised classification methods are applied because the plants present different spectral signatures, depending on two main factors: illumination and the affected state. The colour space used in this application is CIELab, based on the decomposition of the colour in three channels, because it is the closest to human colour perception. The histogram of each channel is successively split into regions by thresholding. The best threshold to be applied is automatically obtained as a combination of three thresholding strategies: (a) Otsu’s method, (b) Isodata algorithm, and (c) Fuzzy thresholding. The fusion of these automatic thresholding techniques and the design of the classification strategy are some of the main findings of the paper, which allows an estimation of the damages and a prediction of the oat production. PMID:22163940
Deep SOMs for automated feature extraction and classification from big data streaming

NASA Astrophysics Data System (ADS)

Sakkari, Mohamed; Ejbali, Ridha; Zaied, Mourad

2017-03-01

In this paper, we proposed a deep self-organizing map model (Deep-SOMs) for automated features extracting and learning from big data streaming which we benefit from the framework Spark for real time streams and highly parallel data processing. The SOMs deep architecture is based on the notion of abstraction (patterns automatically extract from the raw data, from the less to more abstract). The proposed model consists of three hidden self-organizing layers, an input and an output layer. Each layer is made up of a multitude of SOMs, each map only focusing at local headmistress sub-region from the input image. Then, each layer trains the local information to generate more overall information in the higher layer. The proposed Deep-SOMs model is unique in terms of the layers architecture, the SOMs sampling method and learning. During the learning stage we use a set of unsupervised SOMs for feature extraction. We validate the effectiveness of our approach on large data sets such as Leukemia dataset and SRBCT. Results of comparison have shown that the Deep-SOMs model performs better than many existing algorithms for images classification.
Digital image sensor-based assessment of the status of oat (Avena sativa L.) crops after frost damage.

PubMed

Macedo-Cruz, Antonia; Pajares, Gonzalo; Santos, Matilde; Villegas-Romero, Isidro

2011-01-01

The aim of this paper is to classify the land covered with oat crops, and the quantification of frost damage on oats, while plants are still in the flowering stage. The images are taken by a digital colour camera CCD-based sensor. Unsupervised classification methods are applied because the plants present different spectral signatures, depending on two main factors: illumination and the affected state. The colour space used in this application is CIELab, based on the decomposition of the colour in three channels, because it is the closest to human colour perception. The histogram of each channel is successively split into regions by thresholding. The best threshold to be applied is automatically obtained as a combination of three thresholding strategies: (a) Otsu's method, (b) Isodata algorithm, and (c) Fuzzy thresholding. The fusion of these automatic thresholding techniques and the design of the classification strategy are some of the main findings of the paper, which allows an estimation of the damages and a prediction of the oat production.
Nonlinear dimension reduction and clustering by Minimum Curvilinearity unfold neuropathic pain and tissue embryological classes.

PubMed

Cannistraci, Carlo Vittorio; Ravasi, Timothy; Montevecchi, Franco Maria; Ideker, Trey; Alessio, Massimo

2010-09-15

Nonlinear small datasets, which are characterized by low numbers of samples and very high numbers of measures, occur frequently in computational biology, and pose problems in their investigation. Unsupervised hybrid-two-phase (H2P) procedures-specifically dimension reduction (DR), coupled with clustering-provide valuable assistance, not only for unsupervised data classification, but also for visualization of the patterns hidden in high-dimensional feature space. 'Minimum Curvilinearity' (MC) is a principle that-for small datasets-suggests the approximation of curvilinear sample distances in the feature space by pair-wise distances over their minimum spanning tree (MST), and thus avoids the introduction of any tuning parameter. MC is used to design two novel forms of nonlinear machine learning (NML): Minimum Curvilinear embedding (MCE) for DR, and Minimum Curvilinear affinity propagation (MCAP) for clustering. Compared with several other unsupervised and supervised algorithms, MCE and MCAP, whether individually or combined in H2P, overcome the limits of classical approaches. High performance was attained in the visualization and classification of: (i) pain patients (proteomic measurements) in peripheral neuropathy; (ii) human organ tissues (genomic transcription factor measurements) on the basis of their embryological origin. MC provides a valuable framework to estimate nonlinear distances in small datasets. Its extension to large datasets is prefigured for novel NMLs. Classification of neuropathic pain by proteomic profiles offers new insights for future molecular and systems biology characterization of pain. Improvements in tissue embryological classification refine results obtained in an earlier study, and suggest a possible reinterpretation of skin attribution as mesodermal. https://sites.google.com/site/carlovittoriocannistraci/home.
Machine learning in APOGEE. Unsupervised spectral classification with K-means

NASA Astrophysics Data System (ADS)

Garcia-Dias, Rafael; Allende Prieto, Carlos; Sánchez Almeida, Jorge; Ordovás-Pascual, Ignacio

2018-05-01

Context. The volume of data generated by astronomical surveys is growing rapidly. Traditional analysis techniques in spectroscopy either demand intensive human interaction or are computationally expensive. In this scenario, machine learning, and unsupervised clustering algorithms in particular, offer interesting alternatives. The Apache Point Observatory Galactic Evolution Experiment (APOGEE) offers a vast data set of near-infrared stellar spectra, which is perfect for testing such alternatives. Aims: Our research applies an unsupervised classification scheme based on K-means to the massive APOGEE data set. We explore whether the data are amenable to classification into discrete classes. Methods: We apply the K-means algorithm to 153 847 high resolution spectra (R ≈ 22 500). We discuss the main virtues and weaknesses of the algorithm, as well as our choice of parameters. Results: We show that a classification based on normalised spectra captures the variations in stellar atmospheric parameters, chemical abundances, and rotational velocity, among other factors. The algorithm is able to separate the bulge and halo populations, and distinguish dwarfs, sub-giants, RC, and RGB stars. However, a discrete classification in flux space does not result in a neat organisation in the parameters' space. Furthermore, the lack of obvious groups in flux space causes the results to be fairly sensitive to the initialisation, and disrupts the efficiency of commonly-used methods to select the optimal number of clusters. Our classification is publicly available, including extensive online material associated with the APOGEE Data Release 12 (DR12). Conclusions: Our description of the APOGEE database can help greatly with the identification of specific types of targets for various applications. We find a lack of obvious groups in flux space, and identify limitations of the K-means algorithm in dealing with this kind of data. Full Tables B.1-B.4 are only available at the CDS via anonymous ftp to http://cdsarc.u-strasbg.fr (http://130.79.128.5) or via http://cdsarc.u-strasbg.fr/viz-bin/qcat?J/A+A/612/A98
Producing a satellite-derived map and modelling Spartina alterniflora expansion for Willapa Bay in Washington State

NASA Astrophysics Data System (ADS)

Berlin, Cynthia Jane

1998-12-01

This research addresses the identification of the areal extent of the intertidal wetlands of Willapa Bay, Washington, and the evaluation of the potential for exotic Spartina alterniflora (smooth cordgrass) expansion in the bay using a spatial geographic approach. It is hoped that the results will address not only the management needs of the study area but provide a research design that may be applied to studies of other coastal wetlands. Four satellite images, three Landsat Multi-Spectral (MSS) and one Thematic Mapper (TM), are used to derive a map showing areas of water, low, middle and high intertidal, and upland. Two multi-date remote sensing mapping techniques are assessed: a supervised classification using density-slicing and an unsupervised classification using an ISODATA algorithm. Statistical comparisons are made between the resultant derived maps and the U.S.G.S. topographic maps for the Willapa Bay area. The potential for Spartina expansion in the bay is assessed using a sigmoidal (logistic) growth model and a spatial modelling procedure for four possible growth scenarios: without management controls (Business-as-Usual), with moderate management controls (e.g. harvesting to eliminate seed setting), under a hypothetical increase in the growth rate that may reflect favorable environmental changes, and under a hypothetical decrease in the growth rate that may reflect aggressive management controls. Comparisons for the statistics of the two mapping techniques suggest that although the unsupervised classification method performed satisfactorily, the supervised classification (density-slicing) method provided more satisfactory results. Results from the modelling of potential Spartina expansion suggest that Spartina expansion will proceed rapidly for the Business-as-Usual and hypothetical increase in the growth rate scenario, and at a slower rate for the elimination of seed setting and hypothetical decrease in the growth rate scenarios, until all potential habitat is filled.
A Novel Unsupervised Segmentation Quality Evaluation Method for Remote Sensing Images

PubMed Central

Tang, Yunwei; Jing, Linhai; Ding, Haifeng

2017-01-01

The segmentation of a high spatial resolution remote sensing image is a critical step in geographic object-based image analysis (GEOBIA). Evaluating the performance of segmentation without ground truth data, i.e., unsupervised evaluation, is important for the comparison of segmentation algorithms and the automatic selection of optimal parameters. This unsupervised strategy currently faces several challenges in practice, such as difficulties in designing effective indicators and limitations of the spectral values in the feature representation. This study proposes a novel unsupervised evaluation method to quantitatively measure the quality of segmentation results to overcome these problems. In this method, multiple spectral and spatial features of images are first extracted simultaneously and then integrated into a feature set to improve the quality of the feature representation of ground objects. The indicators designed for spatial stratified heterogeneity and spatial autocorrelation are included to estimate the properties of the segments in this integrated feature set. These two indicators are then combined into a global assessment metric as the final quality score. The trade-offs of the combined indicators are accounted for using a strategy based on the Mahalanobis distance, which can be exhibited geometrically. The method is tested on two segmentation algorithms and three testing images. The proposed method is compared with two existing unsupervised methods and a supervised method to confirm its capabilities. Through comparison and visual analysis, the results verified the effectiveness of the proposed method and demonstrated the reliability and improvements of this method with respect to other methods. PMID:29064416
Centered Kernel Alignment Enhancing Neural Network Pretraining for MRI-Based Dementia Diagnosis

PubMed Central

Cárdenas-Peña, David; Collazos-Huertas, Diego; Castellanos-Dominguez, German

2016-01-01

Dementia is a growing problem that affects elderly people worldwide. More accurate evaluation of dementia diagnosis can help during the medical examination. Several methods for computer-aided dementia diagnosis have been proposed using resonance imaging scans to discriminate between patients with Alzheimer's disease (AD) or mild cognitive impairment (MCI) and healthy controls (NC). Nonetheless, the computer-aided diagnosis is especially challenging because of the heterogeneous and intermediate nature of MCI. We address the automated dementia diagnosis by introducing a novel supervised pretraining approach that takes advantage of the artificial neural network (ANN) for complex classification tasks. The proposal initializes an ANN based on linear projections to achieve more discriminating spaces. Such projections are estimated by maximizing the centered kernel alignment criterion that assesses the affinity between the resonance imaging data kernel matrix and the label target matrix. As a result, the performed linear embedding allows accounting for features that contribute the most to the MCI class discrimination. We compare the supervised pretraining approach to two unsupervised initialization methods (autoencoders and Principal Component Analysis) and against the best four performing classification methods of the 2014 CADDementia challenge. As a result, our proposal outperforms all the baselines (7% of classification accuracy and area under the receiver-operating-characteristic curve) at the time it reduces the class biasing. PMID:27148392
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning

PubMed Central

Hoo-Chang, Shin; Roth, Holger R.; Gao, Mingchen; Lu, Le; Xu, Ziyue; Nogues, Isabella; Yao, Jianhua; Mollura, Daniel

2016-01-01

Remarkable progress has been made in image recognition, primarily due to the availability of large-scale annotated datasets (i.e. ImageNet) and the revival of deep convolutional neural networks (CNN). CNNs enable learning data-driven, highly representative, layered hierarchical image features from sufficient training data. However, obtaining datasets as comprehensively annotated as ImageNet in the medical imaging domain remains a challenge. There are currently three major techniques that successfully employ CNNs to medical image classification: training the CNN from scratch, using off-the-shelf pre-trained CNN features, and conducting unsupervised CNN pre-training with supervised fine-tuning. Another effective method is transfer learning, i.e., fine-tuning CNN models (supervised) pre-trained from natural image dataset to medical image tasks (although domain transfer between two medical image datasets is also possible). In this paper, we exploit three important, but previously understudied factors of employing deep convolutional neural networks to computer-aided detection problems. We first explore and evaluate different CNN architectures. The studied models contain 5 thousand to 160 million parameters, and vary in numbers of layers. We then evaluate the influence of dataset scale and spatial image context on performance. Finally, we examine when and why transfer learning from pre-trained ImageNet (via fine-tuning) can be useful. We study two specific computeraided detection (CADe) problems, namely thoraco-abdominal lymph node (LN) detection and interstitial lung disease (ILD) classification. We achieve the state-of-the-art performance on the mediastinal LN detection, with 85% sensitivity at 3 false positive per patient, and report the first five-fold cross-validation classification results on predicting axial CT slices with ILD categories. Our extensive empirical evaluation, CNN model analysis and valuable insights can be extended to the design of high performance CAD systems for other medical imaging tasks. PMID:26886976
Identifying local structural states in atomic imaging by computer vision

DOE Office of Scientific and Technical Information (OSTI.GOV)

Laanait, Nouamane; Ziatdinov, Maxim; He, Qian

The availability of atomically resolved imaging modalities enables an unprecedented view into the local structural states of materials, which manifest themselves by deviations from the fundamental assumptions of periodicity and symmetry. Consequently, approaches that aim to extract these local structural states from atomic imaging data with minimal assumptions regarding the average crystallographic configuration of a material are indispensable to advances in structural and chemical investigations of materials. Here, we present an approach to identify and classify local structural states that is rooted in computer vision. This approach introduces a definition of a structural state that is composed of both localmore » and non-local information extracted from atomically resolved images, and is wholly untethered from the familiar concepts of symmetry and periodicity. Instead, this approach relies on computer vision techniques such as feature detection, and concepts such as scale-invariance. We present the fundamental aspects of local structural state extraction and classification by application to simulated scanning transmission electron microscopy images, and analyze the robustness of this approach in the presence of common instrumental factors such as noise, limited spatial resolution, and weak contrast. Finally, we apply this computer vision-based approach for the unsupervised detection and classification of local structural states in an experimental electron micrograph of a complex oxides interface, and a scanning tunneling micrograph of a defect engineered multilayer graphene surface.« less

Identifying local structural states in atomic imaging by computer vision

DOE PAGES

Laanait, Nouamane; Ziatdinov, Maxim; He, Qian; ...

2016-11-02

The availability of atomically resolved imaging modalities enables an unprecedented view into the local structural states of materials, which manifest themselves by deviations from the fundamental assumptions of periodicity and symmetry. Consequently, approaches that aim to extract these local structural states from atomic imaging data with minimal assumptions regarding the average crystallographic configuration of a material are indispensable to advances in structural and chemical investigations of materials. Here, we present an approach to identify and classify local structural states that is rooted in computer vision. This approach introduces a definition of a structural state that is composed of both localmore » and non-local information extracted from atomically resolved images, and is wholly untethered from the familiar concepts of symmetry and periodicity. Instead, this approach relies on computer vision techniques such as feature detection, and concepts such as scale-invariance. We present the fundamental aspects of local structural state extraction and classification by application to simulated scanning transmission electron microscopy images, and analyze the robustness of this approach in the presence of common instrumental factors such as noise, limited spatial resolution, and weak contrast. Finally, we apply this computer vision-based approach for the unsupervised detection and classification of local structural states in an experimental electron micrograph of a complex oxides interface, and a scanning tunneling micrograph of a defect engineered multilayer graphene surface.« less
Imaging and machine learning techniques for diagnosis of Alzheimer's disease.

PubMed

Mirzaei, Golrokh; Adeli, Anahita; Adeli, Hojjat

2016-12-01

Alzheimer's disease (AD) is a common health problem in elderly people. There has been considerable research toward the diagnosis and early detection of this disease in the past decade. The sensitivity of biomarkers and the accuracy of the detection techniques have been defined to be the key to an accurate diagnosis. This paper presents a state-of-the-art review of the research performed on the diagnosis of AD based on imaging and machine learning techniques. Different segmentation and machine learning techniques used for the diagnosis of AD are reviewed including thresholding, supervised and unsupervised learning, probabilistic techniques, Atlas-based approaches, and fusion of different image modalities. More recent and powerful classification techniques such as the enhanced probabilistic neural network of Ahmadlou and Adeli should be investigated with the goal of improving the diagnosis accuracy. A combination of different image modalities can help improve the diagnosis accuracy rate. Research is needed on the combination of modalities to discover multi-modal biomarkers.
Deep Unfolding for Topic Models.

PubMed

Chien, Jen-Tzung; Lee, Chao-Hsi

2018-02-01

Deep unfolding provides an approach to integrate the probabilistic generative models and the deterministic neural networks. Such an approach is benefited by deep representation, easy interpretation, flexible learning and stochastic modeling. This study develops the unsupervised and supervised learning of deep unfolded topic models for document representation and classification. Conventionally, the unsupervised and supervised topic models are inferred via the variational inference algorithm where the model parameters are estimated by maximizing the lower bound of logarithm of marginal likelihood using input documents without and with class labels, respectively. The representation capability or classification accuracy is constrained by the variational lower bound and the tied model parameters across inference procedure. This paper aims to relax these constraints by directly maximizing the end performance criterion and continuously untying the parameters in learning process via deep unfolding inference (DUI). The inference procedure is treated as the layer-wise learning in a deep neural network. The end performance is iteratively improved by using the estimated topic parameters according to the exponentiated updates. Deep learning of topic models is therefore implemented through a back-propagation procedure. Experimental results show the merits of DUI with increasing number of layers compared with variational inference in unsupervised as well as supervised topic models.
Classification of ROTSE Variable Stars using Machine Learning

NASA Astrophysics Data System (ADS)

Wozniak, P. R.; Akerlof, C.; Amrose, S.; Brumby, S.; Casperson, D.; Gisler, G.; Kehoe, R.; Lee, B.; Marshall, S.; McGowan, K. E.; McKay, T.; Perkins, S.; Priedhorsky, W.; Rykoff, E.; Smith, D. A.; Theiler, J.; Vestrand, W. T.; Wren, J.; ROTSE Collaboration

2001-12-01

We evaluate several Machine Learning algorithms as potential tools for automated classification of variable stars. Using the ROTSE sample of ~1800 variables from a pilot study of 5% of the whole sky, we compare the effectiveness of a supervised technique (Support Vector Machines, SVM) versus unsupervised methods (K-means and Autoclass). There are 8 types of variables in the sample: RR Lyr AB, RR Lyr C, Delta Scuti, Cepheids, detached eclipsing binaries, contact binaries, Miras and LPVs. Preliminary results suggest a very high ( ~95%) efficiency of SVM in isolating a few best defined classes against the rest of the sample, and good accuracy ( ~70-75%) for all classes considered simultaneously. This includes some degeneracies, irreducible with the information at hand. Supervised methods naturally outperform unsupervised methods, in terms of final error rate, but unsupervised methods offer many advantages for large sets of unlabeled data. Therefore, both types of methods should be considered as promising tools for mining vast variability surveys. We project that there are more than 30,000 periodic variables in the ROTSE-I data base covering the entire local sky between V=10 and 15.5 mag. This sample size is already stretching the time capabilities of human analysts.
Exploiting unsupervised and supervised classification for segmentation of the pathological lung in CT

NASA Astrophysics Data System (ADS)

Korfiatis, P.; Kalogeropoulou, C.; Daoussis, D.; Petsas, T.; Adonopoulos, A.; Costaridou, L.

2009-07-01

Delineation of lung fields in presence of diffuse lung diseases (DLPDs), such as interstitial pneumonias (IP), challenges segmentation algorithms. To deal with IP patterns affecting the lung border an automated image texture classification scheme is proposed. The proposed segmentation scheme is based on supervised texture classification between lung tissue (normal and abnormal) and surrounding tissue (pleura and thoracic wall) in the lung border region. This region is coarsely defined around an initial estimate of lung border, provided by means of Markov Radom Field modeling and morphological operations. Subsequently, a support vector machine classifier was trained to distinguish between the above two classes of tissue, using textural feature of gray scale and wavelet domains. 17 patients diagnosed with IP, secondary to connective tissue diseases were examined. Segmentation performance in terms of overlap was 0.924±0.021, and for shape differentiation mean, rms and maximum distance were 1.663±0.816, 2.334±1.574 and 8.0515±6.549 mm, respectively. An accurate, automated scheme is proposed for segmenting abnormal lung fields in HRC affected by IP
Discriminative clustering on manifold for adaptive transductive classification.

PubMed

Zhang, Zhao; Jia, Lei; Zhang, Min; Li, Bing; Zhang, Li; Li, Fanzhang

2017-10-01

In this paper, we mainly propose a novel adaptive transductive label propagation approach by joint discriminative clustering on manifolds for representing and classifying high-dimensional data. Our framework seamlessly combines the unsupervised manifold learning, discriminative clustering and adaptive classification into a unified model. Also, our method incorporates the adaptive graph weight construction with label propagation. Specifically, our method is capable of propagating label information using adaptive weights over low-dimensional manifold features, which is different from most existing studies that usually predict the labels and construct the weights in the original Euclidean space. For transductive classification by our formulation, we first perform the joint discriminative K-means clustering and manifold learning to capture the low-dimensional nonlinear manifolds. Then, we construct the adaptive weights over the learnt manifold features, where the adaptive weights are calculated through performing the joint minimization of the reconstruction errors over features and soft labels so that the graph weights can be joint-optimal for data representation and classification. Using the adaptive weights, we can easily estimate the unknown labels of samples. After that, our method returns the updated weights for further updating the manifold features. Extensive simulations on image classification and segmentation show that our proposed algorithm can deliver the state-of-the-art performance on several public datasets. Copyright © 2017 Elsevier Ltd. All rights reserved.
Preliminary results of the comparative study between EO-1/Hyperion and ALOS/PALSAR

NASA Astrophysics Data System (ADS)

Koizumi, E.; Furuta, R.; Yamamoto, A.

2011-12-01

[Introduction]Hyper-spectral remote sensing images have been used for land-cover classification due to their high spectral resolutions. Synthetic Aperture Radar (SAR) remote sensing data are also useful to probe surface condition because radar image reflects surface geometry, although there are not so many reports about the land-cover detection with combination use of both hyper-spectral data and SAR data. Among SAR sensors, L-band SAR is thought to be useful tool to find physical properties because its comparatively long wave length can through small objects on surface. We are comparing the result of land cover classification and/or physical values from hyper-spectral and L-band SAR data to find the relationship between these two quite different sensors and to confirm the possibility of the combined analysis of hyper-spectral and L-band SAR data, and in this presentation we will report the preliminary result of this study. There are only few sources of both hyper-spectral and L-band SAR data from the space in this time, however, several space organizations plan to launch new satellites on which hyper-spectral or L-band SAR equipments are mounted in next few years. So, the importance of the combined analysis will increase more than ever. [Target Area]We are performing and planning analyses on the following areas in this study. (a)South of Cairo, Nile river area, Egypt, for sand, sandstone, limestone, river, crops. (b)Mount Sakurajima, Japan, for igneous rock and other related geological property. [Methods and Results]EO-1 Hyperion data are analyzed in this study as hyper-spectral data. The Hyperion equipment has 242 channels but some of them include full noise or have no data. We selected channels for analysis by checking each channel, and select about 150 channels (depend on the area). Before analysis, the atmospheric correction of ATCOR-3 was applied for the selected channels. The corrected data were analyzed by unsupervised classification or principal component analysis (PCA). We also did the unsupervised classification with the several components from PCA. According to the analysis results, several classifications can be extracted for each category (vegetation, sand and rocks, and water). One of the interesting results is that there are a few classes for sand as those of other categories, and these classes seem to reflect artificial and natural surface changes that are some result of excavation or scratching. ALOS PALSAR data are analyzed as L-band SAR data. We selected the Dual Polarization data for each target area. The data were converted to backscattered images, and then calculated some image statistic values. The topographic information also calculates with SAR interferometry technique as reference. Comparing the Hyperion classification results with the result of the calculation of statistic values from PALSAR, there are some areas where relativities seem to be confirmed. To confirm the combined analysis between hyper-spectral and L-band SAR data to detect and classify the surface material, further studies are still required. We will continue to investigate more efficient analytic methods and to examine other functions like the adopted channels, the number of class in classification, the kind of statistic information, and so on, to refine the method.
Advanced Unsupervised Classification Methods to Detect Anomalies on Earthen Levees Using Polarimetric SAR Imagery

PubMed Central

Marapareddy, Ramakalavathi; Aanstoos, James V.; Younan, Nicolas H.

2016-01-01

Fully polarimetric Synthetic Aperture Radar (polSAR) data analysis has wide applications for terrain and ground cover classification. The dynamics of surface and subsurface water events can lead to slope instability resulting in slough slides on earthen levees. Early detection of these anomalies by a remote sensing approach could save time versus direct assessment. We used L-band Synthetic Aperture Radar (SAR) to screen levees for anomalies. SAR technology, due to its high spatial resolution and soil penetration capability, is a good choice for identifying problematic areas on earthen levees. Using the parameters entropy (H), anisotropy (A), alpha (α), and eigenvalues (λ, λ1, λ2, and λ3), we implemented several unsupervised classification algorithms for the identification of anomalies on the levee. The classification techniques applied are H/α, H/A, A/α, Wishart H/α, Wishart H/A/α, and H/α/λ classification algorithms. In this work, the effectiveness of the algorithms was demonstrated using quad-polarimetric L-band SAR imagery from the NASA Jet Propulsion Laboratory’s (JPL’s) Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR). The study area is a section of the lower Mississippi River valley in the Southern USA, where earthen flood control levees are maintained by the US Army Corps of Engineers. PMID:27322270
Object-oriented feature-tracking algorithms for SAR images of the marginal ice zone

NASA Technical Reports Server (NTRS)

Daida, Jason; Samadani, Ramin; Vesecky, John F.

1990-01-01

An unsupervised method that chooses and applies the most appropriate tracking algorithm from among different sea-ice tracking algorithms is reported. In contrast to current unsupervised methods, this method chooses and applies an algorithm by partially examining a sequential image pair to draw inferences about what was examined. Based on these inferences the reported method subsequently chooses which algorithm to apply to specific areas of the image pair where that algorithm should work best.
Change classification in SAR time series: a functional approach

NASA Astrophysics Data System (ADS)

Boldt, Markus; Thiele, Antje; Schulz, Karsten; Hinz, Stefan

2017-10-01

Change detection represents a broad field of research in SAR remote sensing, consisting of many different approaches. Besides the simple recognition of change areas, the analysis of type, category or class of the change areas is at least as important for creating a comprehensive result. Conventional strategies for change classification are based on supervised or unsupervised landuse / landcover classifications. The main drawback of such approaches is that the quality of the classification result directly depends on the selection of training and reference data. Additionally, supervised processing methods require an experienced operator who capably selects the training samples. This training step is not necessary when using unsupervised strategies, but nevertheless meaningful reference data must be available for identifying the resulting classes. Consequently, an experienced operator is indispensable. In this study, an innovative concept for the classification of changes in SAR time series data is proposed. Regarding the drawbacks of traditional strategies given above, it copes without using any training data. Moreover, the method can be applied by an operator, who does not have detailed knowledge about the available scenery yet. This knowledge is provided by the algorithm. The final step of the procedure, which main aspect is given by the iterative optimization of an initial class scheme with respect to the categorized change objects, is represented by the classification of these objects to the finally resulting classes. This assignment step is subject of this paper.
Unsupervised Learning in an Ensemble of Spiking Neural Networks Mediated by ITDP.

PubMed

Shim, Yoonsik; Philippides, Andrew; Staras, Kevin; Husbands, Phil

2016-10-01

We propose a biologically plausible architecture for unsupervised ensemble learning in a population of spiking neural network classifiers. A mixture of experts type organisation is shown to be effective, with the individual classifier outputs combined via a gating network whose operation is driven by input timing dependent plasticity (ITDP). The ITDP gating mechanism is based on recent experimental findings. An abstract, analytically tractable model of the ITDP driven ensemble architecture is derived from a logical model based on the probabilities of neural firing events. A detailed analysis of this model provides insights that allow it to be extended into a full, biologically plausible, computational implementation of the architecture which is demonstrated on a visual classification task. The extended model makes use of a style of spiking network, first introduced as a model of cortical microcircuits, that is capable of Bayesian inference, effectively performing expectation maximization. The unsupervised ensemble learning mechanism, based around such spiking expectation maximization (SEM) networks whose combined outputs are mediated by ITDP, is shown to perform the visual classification task well and to generalize to unseen data. The combined ensemble performance is significantly better than that of the individual classifiers, validating the ensemble architecture and learning mechanisms. The properties of the full model are analysed in the light of extensive experiments with the classification task, including an investigation into the influence of different input feature selection schemes and a comparison with a hierarchical STDP based ensemble architecture.
Unsupervised Learning in an Ensemble of Spiking Neural Networks Mediated by ITDP

PubMed Central

Staras, Kevin

2016-01-01

We propose a biologically plausible architecture for unsupervised ensemble learning in a population of spiking neural network classifiers. A mixture of experts type organisation is shown to be effective, with the individual classifier outputs combined via a gating network whose operation is driven by input timing dependent plasticity (ITDP). The ITDP gating mechanism is based on recent experimental findings. An abstract, analytically tractable model of the ITDP driven ensemble architecture is derived from a logical model based on the probabilities of neural firing events. A detailed analysis of this model provides insights that allow it to be extended into a full, biologically plausible, computational implementation of the architecture which is demonstrated on a visual classification task. The extended model makes use of a style of spiking network, first introduced as a model of cortical microcircuits, that is capable of Bayesian inference, effectively performing expectation maximization. The unsupervised ensemble learning mechanism, based around such spiking expectation maximization (SEM) networks whose combined outputs are mediated by ITDP, is shown to perform the visual classification task well and to generalize to unseen data. The combined ensemble performance is significantly better than that of the individual classifiers, validating the ensemble architecture and learning mechanisms. The properties of the full model are analysed in the light of extensive experiments with the classification task, including an investigation into the influence of different input feature selection schemes and a comparison with a hierarchical STDP based ensemble architecture. PMID:27760125
A comparison of neural network and fuzzy clustering techniques in segmenting magnetic resonance images of the brain

NASA Technical Reports Server (NTRS)

Hall, Lawrence O.; Bensaid, Amine M.; Clarke, Laurence P.; Velthuizen, Robert P.; Silbiger, Martin S.; Bezdek, James C.

1992-01-01

Magnetic resonance (MR) brain section images are segmented and then synthetically colored to give visual representations of the original data with three approaches: the literal and approximate fuzzy c-means unsupervised clustering algorithms and a supervised computational neural network, a dynamic multilayered perception trained with the cascade correlation learning algorithm. Initial clinical results are presented on both normal volunteers and selected patients with brain tumors surrounded by edema. Supervised and unsupervised segmentation techniques provide broadly similar results. Unsupervised fuzzy algorithms were visually observed to show better segmentation when compared with raw image data for volunteer studies. However, for a more complex segmentation problem with tumor/edema or cerebrospinal fluid boundary, where the tissues have similar MR relaxation behavior, inconsistency in rating among experts was observed.
Visualization of suspicious lesions in breast MRI based on intelligent neural systems

NASA Astrophysics Data System (ADS)

Twellmann, Thorsten; Lange, Oliver; Nattkemper, Tim Wilhelm; Meyer-Bäse, Anke

2006-05-01

Intelligent medical systems based on supervised and unsupervised artificial neural networks are applied to the automatic visualization and classification of suspicious lesions in breast MRI. These systems represent an important component of future sophisticated computer-aided diagnosis systems and enable the extraction of spatial and temporal features of dynamic MRI data stemming from patients with confirmed lesion diagnosis. By taking into account the heterogenity of the cancerous tissue, these techniques reveal the malignant, benign and normal kinetic signals and and provide a regional subclassification of pathological breast tissue. Intelligent medical systems are expected to have substantial implications in healthcare politics by contributing to the diagnosis of indeterminate breast lesions by non-invasive imaging.
Near ground level sensing for spatial analysis of vegetation

NASA Technical Reports Server (NTRS)

Sauer, Tom; Rasure, John; Gage, Charlie

1991-01-01

Measured changes in vegetation indicate the dynamics of ecological processes and can identify the impacts from disturbances. Traditional methods of vegetation analysis tend to be slow because they are labor intensive; as a result, these methods are often confined to small local area measurements. Scientists need new algorithms and instruments that will allow them to efficiently study environmental dynamics across a range of different spatial scales. A new methodology that addresses this problem is presented. This methodology includes the acquisition, processing, and presentation of near ground level image data and its corresponding spatial characteristics. The systematic approach taken encompasses a feature extraction process, a supervised and unsupervised classification process, and a region labeling process yielding spatial information.
Automated surface quality inspection with ARGOS: a case study

NASA Astrophysics Data System (ADS)

Kiefhaber, Daniel; Etzold, Fabian; Warken, Arno F.; Asfour, Jean-Michel

2017-06-01

The commercial availability of automated inspection systems for optical surfaces specified according to ISO 10110-7 promises unsupervised and automated quality control with reproducible results. In this study, the classification results of the ARGOS inspection system are compared to the decisions by well-trained inspectors based on manual-visual inspection. Both are found to agree in 93.6% of the studied cases. Exemplary cases with differing results are studied, and shown to be partly caused by shortcomings of the ISO 10110-7 standard, which was written for the industry standard manual-visual inspection. Applying it to high resolution images of the whole surface of objective machine vision systems brings with it a few challenges which are discussed.
Superpixel-based classification of gastric chromoendoscopy images

NASA Astrophysics Data System (ADS)

Boschetto, Davide; Grisan, Enrico

2017-03-01

Chromoendoscopy (CH) is a gastroenterology imaging modality that involves the staining of tissues with methylene blue, which reacts with the internal walls of the gastrointestinal tract, improving the visual contrast in mucosal surfaces and thus enhancing a doctor's ability to screen precancerous lesions or early cancer. This technique helps identify areas that can be targeted for biopsy or treatment and in this work we will focus on gastric cancer detection. Gastric chromoendoscopy for cancer detection has several taxonomies available, one of which classifies CH images into three classes (normal, metaplasia, dysplasia) based on color, shape and regularity of pit patterns. Computer-assisted diagnosis is desirable to help us improve the reliability of the tissue classification and abnormalities detection. However, traditional computer vision methodologies, mainly segmentation, do not translate well to the specific visual characteristics of a gastroenterology imaging scenario. We propose the exploitation of a first unsupervised segmentation via superpixel, which groups pixels into perceptually meaningful atomic regions, used to replace the rigid structure of the pixel grid. For each superpixel, a set of features is extracted and then fed to a random forest based classifier, which computes a model used to predict the class of each superpixel. The average general accuracy of our model is 92.05% in the pixel domain (86.62% in the superpixel domain), while detection accuracies on the normal and abnormal class are respectively 85.71% and 95%. Eventually, the whole image class can be predicted image through a majority vote on each superpixel's predicted class.
A Hessian-based methodology for automatic surface crack detection and classification from pavement images

NASA Astrophysics Data System (ADS)

Ghanta, Sindhu; Shahini Shamsabadi, Salar; Dy, Jennifer; Wang, Ming; Birken, Ralf

2015-04-01

Around 3,000,000 million vehicle miles are annually traveled utilizing the US transportation systems alone. In addition to the road traffic safety, maintaining the road infrastructure in a sound condition promotes a more productive and competitive economy. Due to the significant amounts of financial and human resources required to detect surface cracks by visual inspection, detection of these surface defects are often delayed resulting in deferred maintenance operations. This paper introduces an automatic system for acquisition, detection, classification, and evaluation of pavement surface cracks by unsupervised analysis of images collected from a camera mounted on the rear of a moving vehicle. A Hessian-based multi-scale filter has been utilized to detect ridges in these images at various scales. Post-processing on the extracted features has been implemented to produce statistics of length, width, and area covered by cracks, which are crucial for roadway agencies to assess pavement quality. This process has been realized on three sets of roads with different pavement conditions in the city of Brockton, MA. A ground truth dataset labeled manually is made available to evaluate this algorithm and results rendered more than 90% segmentation accuracy demonstrating the feasibility of employing this approach at a larger scale.
Unsupervised learning on scientific ocean drilling datasets from the South China Sea

NASA Astrophysics Data System (ADS)

Tse, Kevin C.; Chiu, Hon-Chim; Tsang, Man-Yin; Li, Yiliang; Lam, Edmund Y.

2018-06-01

Unsupervised learning methods were applied to explore data patterns in multivariate geophysical datasets collected from ocean floor sediment core samples coming from scientific ocean drilling in the South China Sea. Compared to studies on similar datasets, but using supervised learning methods which are designed to make predictions based on sample training data, unsupervised learning methods require no a priori information and focus only on the input data. In this study, popular unsupervised learning methods including K-means, self-organizing maps, hierarchical clustering and random forest were coupled with different distance metrics to form exploratory data clusters. The resulting data clusters were externally validated with lithologic units and geologic time scales assigned to the datasets by conventional methods. Compact and connected data clusters displayed varying degrees of correspondence with existing classification by lithologic units and geologic time scales. K-means and self-organizing maps were observed to perform better with lithologic units while random forest corresponded best with geologic time scales. This study sets a pioneering example of how unsupervised machine learning methods can be used as an automatic processing tool for the increasingly high volume of scientific ocean drilling data.
A learning tool for optical and microwave satellite image processing and analysis

NASA Astrophysics Data System (ADS)

Dashondhi, Gaurav K.; Mohanty, Jyotirmoy; Eeti, Laxmi N.; Bhattacharya, Avik; De, Shaunak; Buddhiraju, Krishna M.

2016-04-01

This paper presents a self-learning tool, which contains a number of virtual experiments for processing and analysis of Optical/Infrared and Synthetic Aperture Radar (SAR) images. The tool is named Virtual Satellite Image Processing and Analysis Lab (v-SIPLAB) Experiments that are included in Learning Tool are related to: Optical/Infrared - Image and Edge enhancement, smoothing, PCT, vegetation indices, Mathematical Morphology, Accuracy Assessment, Supervised/Unsupervised classification etc.; Basic SAR - Parameter extraction and range spectrum estimation, Range compression, Doppler centroid estimation, Azimuth reference function generation and compression, Multilooking, image enhancement, texture analysis, edge and detection. etc.; SAR Interferometry - BaseLine Calculation, Extraction of single look SAR images, Registration, Resampling, and Interferogram generation; SAR Polarimetry - Conversion of AirSAR or Radarsat data to S2/C3/T3 matrix, Speckle Filtering, Power/Intensity image generation, Decomposition of S2/C3/T3, Classification of S2/C3/T3 using Wishart Classifier [3]. A professional quality polarimetric SAR software can be found at [8], a part of whose functionality can be found in our system. The learning tool also contains other modules, besides executable software experiments, such as aim, theory, procedure, interpretation, quizzes, link to additional reading material and user feedback. Students can have understanding of Optical and SAR remotely sensed images through discussion of basic principles and supported by structured procedure for running and interpreting the experiments. Quizzes for self-assessment and a provision for online feedback are also being provided to make this Learning tool self-contained. One can download results after performing experiments.

Geological applications of machine learning on hyperspectral remote sensing data

NASA Astrophysics Data System (ADS)

Tse, C. H.; Li, Yi-liang; Lam, Edmund Y.

2015-02-01

The CRISM imaging spectrometer orbiting Mars has been producing a vast amount of data in the visible to infrared wavelengths in the form of hyperspectral data cubes. These data, compared with those obtained from previous remote sensing techniques, yield an unprecedented level of detailed spectral resolution in additional to an ever increasing level of spatial information. A major challenge brought about by the data is the burden of processing and interpreting these datasets and extract the relevant information from it. This research aims at approaching the challenge by exploring machine learning methods especially unsupervised learning to achieve cluster density estimation and classification, and ultimately devising an efficient means leading to identification of minerals. A set of software tools have been constructed by Python to access and experiment with CRISM hyperspectral cubes selected from two specific Mars locations. A machine learning pipeline is proposed and unsupervised learning methods were implemented onto pre-processed datasets. The resulting data clusters are compared with the published ASTER spectral library and browse data products from the Planetary Data System (PDS). The result demonstrated that this approach is capable of processing the huge amount of hyperspectral data and potentially providing guidance to scientists for more detailed studies.
Space Object Classification Using Fused Features of Time Series Data

NASA Astrophysics Data System (ADS)

Jia, B.; Pham, K. D.; Blasch, E.; Shen, D.; Wang, Z.; Chen, G.

In this paper, a fused feature vector consisting of raw time series and texture feature information is proposed for space object classification. The time series data includes historical orbit trajectories and asteroid light curves. The texture feature is derived from recurrence plots using Gabor filters for both unsupervised learning and supervised learning algorithms. The simulation results show that the classification algorithms using the fused feature vector achieve better performance than those using raw time series or texture features only.
A Novel Semi-Supervised Methodology for Extracting Tumor Type-Specific MRS Sources in Human Brain Data

PubMed Central

Ortega-Martorell, Sandra; Ruiz, Héctor; Vellido, Alfredo; Olier, Iván; Romero, Enrique; Julià-Sapé, Margarida; Martín, José D.; Jarman, Ian H.; Arús, Carles; Lisboa, Paulo J. G.

2013-01-01

Background The clinical investigation of human brain tumors often starts with a non-invasive imaging study, providing information about the tumor extent and location, but little insight into the biochemistry of the analyzed tissue. Magnetic Resonance Spectroscopy can complement imaging by supplying a metabolic fingerprint of the tissue. This study analyzes single-voxel magnetic resonance spectra, which represent signal information in the frequency domain. Given that a single voxel may contain a heterogeneous mix of tissues, signal source identification is a relevant challenge for the problem of tumor type classification from the spectroscopic signal. Methodology/Principal Findings Non-negative matrix factorization techniques have recently shown their potential for the identification of meaningful sources from brain tissue spectroscopy data. In this study, we use a convex variant of these methods that is capable of handling negatively-valued data and generating sources that can be interpreted as tumor class prototypes. A novel approach to convex non-negative matrix factorization is proposed, in which prior knowledge about class information is utilized in model optimization. Class-specific information is integrated into this semi-supervised process by setting the metric of a latent variable space where the matrix factorization is carried out. The reported experimental study comprises 196 cases from different tumor types drawn from two international, multi-center databases. The results indicate that the proposed approach outperforms a purely unsupervised process by achieving near perfect correlation of the extracted sources with the mean spectra of the tumor types. It also improves tissue type classification. Conclusions/Significance We show that source extraction by unsupervised matrix factorization benefits from the integration of the available class information, so operating in a semi-supervised learning manner, for discriminative source identification and brain tumor labeling from single-voxel spectroscopy data. We are confident that the proposed methodology has wider applicability for biomedical signal processing. PMID:24376744
Quantitative radiomic profiling of glioblastoma represents transcriptomic expression.

PubMed

Kong, Doo-Sik; Kim, Junhyung; Ryu, Gyuha; You, Hye-Jin; Sung, Joon Kyung; Han, Yong Hee; Shin, Hye-Mi; Lee, In-Hee; Kim, Sung-Tae; Park, Chul-Kee; Choi, Seung Hong; Choi, Jeong Won; Seol, Ho Jun; Lee, Jung-Il; Nam, Do-Hyun

2018-01-19

Quantitative imaging biomarkers have increasingly emerged in the field of research utilizing available imaging modalities. We aimed to identify good surrogate radiomic features that can represent genetic changes of tumors, thereby establishing noninvasive means for predicting treatment outcome. From May 2012 to June 2014, we retrospectively identified 65 patients with treatment-naïve glioblastoma with available clinical information from the Samsung Medical Center data registry. Preoperative MR imaging data were obtained for all 65 patients with primary glioblastoma. A total of 82 imaging features including first-order statistics, volume, and size features, were semi-automatically extracted from structural and physiologic images such as apparent diffusion coefficient and perfusion images. Using commercially available software, NordicICE, we performed quantitative imaging analysis and collected the dataset composed of radiophenotypic parameters. Unsupervised clustering methods revealed that the radiophenotypic dataset was composed of three clusters. Each cluster represented a distinct molecular classification of glioblastoma; classical type, proneural and neural types, and mesenchymal type. These clusters also reflected differential clinical outcomes. We found that extracted imaging signatures does not represent copy number variation and somatic mutation. Quantitative radiomic features provide a potential evidence to predict molecular phenotype and treatment outcome. Radiomic profiles represents transcriptomic phenotypes more well.
Diagnostic index of 3D osteoarthritic changes in TMJ condylar morphology

NASA Astrophysics Data System (ADS)

Gomes, Liliane R.; Gomes, Marcelo; Jung, Bryan; Paniagua, Beatriz; Ruellas, Antonio C.; Gonçalves, João. Roberto; Styner, Martin A.; Wolford, Larry; Cevidanes, Lucia

2015-03-01

The aim of this study was to investigate imaging statistical approaches for classifying 3D osteoarthritic morphological variations among 169 Temporomandibular Joint (TMJ) condyles. Cone beam Computed Tomography (CBCT) scans were acquired from 69 patients with long-term TMJ Osteoarthritis (OA) (39.1 ± 15.7 years), 15 patients at initial diagnosis of OA (44.9 ± 14.8 years) and 7 healthy controls (43 ± 12.4 years). 3D surface models of the condyles were constructed and Shape Correspondence was used to establish correspondent points on each model. The statistical framework included a multivariate analysis of covariance (MANCOVA) and Direction-Projection- Permutation (DiProPerm) for testing statistical significance of the differences between healthy control and the OA group determined by clinical and radiographic diagnoses. Unsupervised classification using hierarchical agglomerative clustering (HAC) was then conducted. Condylar morphology in OA and healthy subjects varied widely. Compared with healthy controls, OA average condyle was statistically significantly smaller in all dimensions except its anterior surface. Significant flattening of the lateral pole was noticed at initial diagnosis (p < 0.05). It was observed areas of 3.88 mm bone resorption at the superior surface and 3.10 mm bone apposition at the anterior aspect of the long-term OA average model. 1000 permutation statistics of DiProPerm supported a significant difference between the healthy control group and OA group (t = 6.7, empirical p-value = 0.001). Clinically meaningful unsupervised classification of TMJ condylar morphology determined a preliminary diagnostic index of 3D osteoarthritic changes, which may be the first step towards a more targeted diagnosis of this condition.
An unsupervised classification method for inferring original case locations from low-resolution disease maps.

PubMed

Brownstein, John S; Cassa, Christopher A; Kohane, Isaac S; Mandl, Kenneth D

2006-12-08

Widespread availability of geographic information systems software has facilitated the use of disease mapping in academia, government and private sector. Maps that display the address of affected patients are often exchanged in public forums, and published in peer-reviewed journal articles. As previously reported, a search of figure legends in five major medical journals found 19 articles from 1994-2004 that identify over 19,000 patient addresses. In this report, a method is presented to evaluate whether patient privacy is being breached in the publication of low-resolution disease maps. To demonstrate the effect, a hypothetical low-resolution map of geocoded patient addresses was created and the accuracy with which patient addresses can be resolved is described. Through georeferencing and unsupervised classification of the original image, the method precisely re-identified 26% (144/550) of the patient addresses from a presentation quality map and 79% (432/550) from a publication quality map. For the presentation quality map, 99.8% of the addresses were within 70 meters (approximately one city block length) of the predicted patient location, 51.6% of addresses were identified within five buildings, 70.7% within ten buildings and 93% within twenty buildings. For the publication quality map, all addresses were within 14 meters and 11 buildings of the predicted patient location. This study demonstrates that lowering the resolution of a map displaying geocoded patient addresses does not sufficiently protect patient addresses from re-identification. Guidelines to protect patient privacy, including those of medical journals, should reflect policies that ensure privacy protection when spatial data are displayed or published.
Change Detection Analysis of Costal Habitat Using Remote Sensing Technologies in the Western Arabian Gulf (Saudi Arabian Coast) over a Thirty-Year Period.

NASA Astrophysics Data System (ADS)

El-Askary, H. M.; Idris, N.; Johnson, S. H.; Qurban, M. A. B.

2014-12-01

Many factors can severely affect the growth and abundance of the marine ecosystems. For example, due to anthropogenic and natural forces, benthic habitats including but not limited to mangroves, sea grass, salt marshes, macro algae, and coral reefs have been experiencing high levels of declination. Furthermore, aerosols and their propellants are suspected contributors to marine habitat degradation. Although several studies reveal that the Arabian Gulf habitats have suffered deleterious impacts after the Gulf War and the following six month off-shore oil spill, limited research exists to track the changes in benthic habitats over the past three decades using remote sensing. Document changes in costal habitats over the past thirty years were better observed with the use of multispectral remote sensors such as Landsat-5, Landsat-7, and Landsat8 (OLI). Change detection analysis was performed on the three Landsat images (Landsat-5 for the 1987 image, Landsat-7 for the 2000, and Landsat-8 for the 2013 image). The images were then modified, masked off from open water and land. An unsupervised classification was performed which cluster similar classes together. The supervised classification displayed the seven following classes: coral reefs, macro algae, sea grass, salt marshes, mangroves, water, and land. Compared to 1987 image to 2000 scene, there was a noticeable increase in the extensiveness of salt marsh and macro algae habitats. However, a significant decrease in salt marsh habitats were apparent in the 2013 scene.
Unsupervised categorization method of graphemes on handwritten manuscripts: application to style recognition

NASA Astrophysics Data System (ADS)

Daher, H.; Gaceb, D.; Eglin, V.; Bres, S.; Vincent, N.

2012-01-01

We present in this paper a feature selection and weighting method for medieval handwriting images that relies on codebooks of shapes of small strokes of characters (graphemes that are issued from the decomposition of manuscripts). These codebooks are important to simplify the automation of the analysis, the manuscripts transcription and the recognition of styles or writers. Our approach provides a precise features weighting by genetic algorithms and a highperformance methodology for the categorization of the shapes of graphemes by using graph coloring into codebooks which are applied in turn on CBIR (Content Based Image Retrieval) in a mixed handwriting database containing different pages from different writers, periods of the history and quality. We show how the coupling of these two mechanisms 'features weighting - graphemes classification' can offer a better separation of the forms to be categorized by exploiting their grapho-morphological, their density and their significant orientations particularities.
Generalized Wishart Mixtures for Unsupervised Classification of PolSAR Data

NASA Astrophysics Data System (ADS)

Li, Lan; Chen, Erxue; Li, Zengyuan

2013-01-01

This paper presents an unsupervised clustering algorithm based upon the expectation maximization (EM) algorithm for finite mixture modelling, using the complex wishart probability density function (PDF) for the probabilities. The mixture model enables to consider heterogeneous thematic classes which could not be better fitted by the unimodal wishart distribution. In order to make it fast and robust to calculate, we use the recently proposed generalized gamma distribution (GΓD) for the single polarization intensity data to make the initial partition. Then we use the wishart probability density function for the corresponding sample covariance matrix to calculate the posterior class probabilities for each pixel. The posterior class probabilities are used for the prior probability estimates of each class and weights for all class parameter updates. The proposed method is evaluated and compared with the wishart H-Alpha-A classification. Preliminary results show that the proposed method has better performance.
Analysis of thematic mapper simulator data collected over eastern North Dakota

NASA Technical Reports Server (NTRS)

Anderson, J. E. (Principal Investigator)

1982-01-01

The results of the analysis of aircraft-acquired thematic mapper simulator (TMS) data, collected to investigate the utility of thematic mapper data in crop area and land cover estimates, are discussed. Results of the analysis indicate that the seven-channel TMS data are capable of delineating the 13 crop types included in the study to an overall pixel classification accuracy of 80.97% correct, with relative efficiencies for four crop types examined between 1.62 and 26.61. Both supervised and unsupervised spectral signature development techniques were evaluated. The unsupervised methods proved to be inferior (based on analysis of variance) for the majority of crop types considered. Given the ground truth data set used for spectral signature development as well as evaluation of performance, it is possible to demonstrate which signature development technique would produce the highest percent correct classification for each crop type.
Signature extension: An approach to operational multispectral surveys

NASA Technical Reports Server (NTRS)

Nalepka, R. F.; Morgenstern, J. P.

1973-01-01

Two data processing techniques were suggested as applicable to the large area survey problem. One approach was to use unsupervised classification (clustering) techniques. Investigation of this method showed that since the method did nothing to reduce the signal variability, the use of this method would be very time consuming and possibly inaccurate as well. The conclusion is that unsupervised classification techniques of themselves are not a solution to the large area survey problem. The other method investigated was the use of signature extension techniques. Such techniques function by normalizing the data to some reference condition. Thus signatures from an isolated area could be used to process large quantities of data. In this manner, ground information requirements and computer training are minimized. Several signature extension techniques were tested. The best of these allowed signatures to be extended between data sets collected four days and 80 miles apart with an average accuracy of better than 90%.
Tree leaves extraction in natural images: comparative study of preprocessing tools and segmentation methods.

PubMed

Grand-Brochier, Manuel; Vacavant, Antoine; Cerutti, Guillaume; Kurtz, Camille; Weber, Jonathan; Tougne, Laure

2015-05-01

In this paper, we propose a comparative study of various segmentation methods applied to the extraction of tree leaves from natural images. This study follows the design of a mobile application, developed by Cerutti et al. (published in ReVeS Participation--Tree Species Classification Using Random Forests and Botanical Features. CLEF 2012), to highlight the impact of the choices made for segmentation aspects. All the tests are based on a database of 232 images of tree leaves depicted on natural background from smartphones acquisitions. We also propose to study the improvements, in terms of performance, using preprocessing tools, such as the interaction between the user and the application through an input stroke, as well as the use of color distance maps. The results presented in this paper shows that the method developed by Cerutti et al. (denoted Guided Active Contour), obtains the best score for almost all observation criteria. Finally, we detail our online benchmark composed of 14 unsupervised methods and 6 supervised ones.
Automatic Spatio-Temporal Flow Velocity Measurement in Small Rivers Using Thermal Image Sequences

NASA Astrophysics Data System (ADS)

Lin, D.; Eltner, A.; Sardemann, H.; Maas, H.-G.

2018-05-01

An automatic spatio-temporal flow velocity measurement approach, using an uncooled thermal camera, is proposed in this paper. The basic principle of the method is to track visible thermal features at the water surface in thermal camera image sequences. Radiometric and geometric calibrations are firstly implemented to remove vignetting effects in thermal imagery and to get the interior orientation parameters of the camera. An object-based unsupervised classification approach is then applied to detect the interest regions for data referencing and thermal feature tracking. Subsequently, GCPs are extracted to orient the river image sequences and local hot points are identified as tracking features. Afterwards, accurate dense tracking outputs are obtained using pyramidal Lucas-Kanade method. To validate the accuracy potential of the method, measurements obtained from thermal feature tracking are compared with reference measurements taken by a propeller gauge. Results show a great potential of automatic flow velocity measurement in small rivers using imagery from a thermal camera.
Mesoscale monitoring of the soil freeze/thaw boundary from orbital microwave radiometry

NASA Technical Reports Server (NTRS)

Dobson, Craig; Ulaby, Fawwaz T.; Zuerndorfer, Brian; England, Anthony W.

1990-01-01

A technique was developed for mapping the spatial extent of frozen soils from the spectral characteristics of the 10.7 to 37 GHz radiobrightness. Through computational models for the spectral radiobrightness of diurnally heated freesing soils, a distinctive radiobrightness signature was identified for frozen soils, and the signature was cast as a discriminant for unsupervised classification. In addition to large area images, local area spatial averages of radiobrightness were calculated for each radiobrightness channel at 7 meteorologic sites within the test region. Local area averages at the meteorologic sites were used to define the preliminary boundaries in the Freeze Indicator discriminate. Freeze Indicator images based upon Nimbus 7, Scanning Multichannel Microwave Radiometer (SMMR) data effectively map temporal variations in the freeze/thaw pattern for the northern Great Plains at the time scale of days. Diurnal thermal gradients have a small but measurable effect upon the SMMR spectral gradient. Scale-space filtering can be used to improve the spatial resolution of a freeze/thaw classified image.
A Framework for Land Cover Classification Using Discrete Return LiDAR Data: Adopting Pseudo-Waveform and Hierarchical Segmentation

NASA Technical Reports Server (NTRS)

Jung, Jinha; Pasolli, Edoardo; Prasad, Saurabh; Tilton, James C.; Crawford, Melba M.

2014-01-01

Acquiring current, accurate land-use information is critical for monitoring and understanding the impact of anthropogenic activities on natural environments.Remote sensing technologies are of increasing importance because of their capability to acquire information for large areas in a timely manner, enabling decision makers to be more effective in complex environments. Although optical imagery has demonstrated to be successful for land cover classification, active sensors, such as light detection and ranging (LiDAR), have distinct capabilities that can be exploited to improve classification results. However, utilization of LiDAR data for land cover classification has not been fully exploited. Moreover, spatial-spectral classification has recently gained significant attention since classification accuracy can be improved by extracting additional information from the neighboring pixels. Although spatial information has been widely used for spectral data, less attention has been given to LiDARdata. In this work, a new framework for land cover classification using discrete return LiDAR data is proposed. Pseudo-waveforms are generated from the LiDAR data and processed by hierarchical segmentation. Spatial featuresare extracted in a region-based way using a new unsupervised strategy for multiple pruning of the segmentation hierarchy. The proposed framework is validated experimentally on a real dataset acquired in an urban area. Better classification results are exhibited by the proposed framework compared to the cases in which basic LiDAR products such as digital surface model and intensity image are used. Moreover, the proposed region-based feature extraction strategy results in improved classification accuracies in comparison with a more traditional window-based approach.
An information-based network approach for protein classification

PubMed Central

Wan, Xiaogeng; Zhao, Xin; Yau, Stephen S. T.

2017-01-01

Protein classification is one of the critical problems in bioinformatics. Early studies used geometric distances and polygenetic-tree to classify proteins. These methods use binary trees to present protein classification. In this paper, we propose a new protein classification method, whereby theories of information and networks are used to classify the multivariate relationships of proteins. In this study, protein universe is modeled as an undirected network, where proteins are classified according to their connections. Our method is unsupervised, multivariate, and alignment-free. It can be applied to the classification of both protein sequences and structures. Nine examples are used to demonstrate the efficiency of our new method. PMID:28350835
Automated cell-type classification in intact tissues by single-cell molecular profiling

PubMed Central

2018-01-01

A major challenge in biology is identifying distinct cell classes and mapping their interactions in vivo. Tissue-dissociative technologies enable deep single cell molecular profiling but do not provide spatial information. We developed a proximity ligation in situ hybridization technology (PLISH) with exceptional signal strength, specificity, and sensitivity in tissue. Multiplexed data sets can be acquired using barcoded probes and rapid label-image-erase cycles, with automated calculation of single cell profiles, enabling clustering and anatomical re-mapping of cells. We apply PLISH to expression profile ~2900 cells in intact mouse lung, which identifies and localizes known cell types, including rare ones. Unsupervised classification of the cells indicates differential expression of ‘housekeeping’ genes between cell types, and re-mapping of two sub-classes of Club cells highlights their segregated spatial domains in terminal airways. By enabling single cell profiling of various RNA species in situ, PLISH can impact many areas of basic and medical research. PMID:29319504
Mapping the Philippines' mangrove forests using Landsat imagery

USGS Publications Warehouse

Long, Jordan; Giri, Chandra

2011-01-01

Current, accurate, and reliable information on the areal extent and spatial distribution of mangrove forests in the Philippines is limited. Previous estimates of mangrove extent do not illustrate the spatial distribution for the entire country. This study, part of a global assessment of mangrove dynamics, mapped the spatial distribution and areal extent of the Philippines’ mangroves circa 2000. We used publicly available Landsat data acquired primarily from the Global Land Survey to map the total extent and spatial distribution. ISODATA clustering, an unsupervised classification technique, was applied to 61 Landsat images. Statistical analysis indicates the total area of mangrove forest cover was approximately 256,185 hectares circa 2000 with overall classification accuracy of 96.6% and a kappa coefficient of 0.926. These results differ substantially from most recent estimates of mangrove area in the Philippines. The results of this study may assist the decision making processes for rehabilitation and conservation efforts that are currently needed to protect and restore the Philippines’ degraded mangrove forests.
Hierarchical Gene Selection and Genetic Fuzzy System for Cancer Microarray Data Classification

PubMed Central

Nguyen, Thanh; Khosravi, Abbas; Creighton, Douglas; Nahavandi, Saeid

2015-01-01

This paper introduces a novel approach to gene selection based on a substantial modification of analytic hierarchy process (AHP). The modified AHP systematically integrates outcomes of individual filter methods to select the most informative genes for microarray classification. Five individual ranking methods including t-test, entropy, receiver operating characteristic (ROC) curve, Wilcoxon and signal to noise ratio are employed to rank genes. These ranked genes are then considered as inputs for the modified AHP. Additionally, a method that uses fuzzy standard additive model (FSAM) for cancer classification based on genes selected by AHP is also proposed in this paper. Traditional FSAM learning is a hybrid process comprising unsupervised structure learning and supervised parameter tuning. Genetic algorithm (GA) is incorporated in-between unsupervised and supervised training to optimize the number of fuzzy rules. The integration of GA enables FSAM to deal with the high-dimensional-low-sample nature of microarray data and thus enhance the efficiency of the classification. Experiments are carried out on numerous microarray datasets. Results demonstrate the performance dominance of the AHP-based gene selection against the single ranking methods. Furthermore, the combination of AHP-FSAM shows a great accuracy in microarray data classification compared to various competing classifiers. The proposed approach therefore is useful for medical practitioners and clinicians as a decision support system that can be implemented in the real medical practice. PMID:25823003
Hierarchical gene selection and genetic fuzzy system for cancer microarray data classification.

PubMed

Nguyen, Thanh; Khosravi, Abbas; Creighton, Douglas; Nahavandi, Saeid

2015-01-01

This paper introduces a novel approach to gene selection based on a substantial modification of analytic hierarchy process (AHP). The modified AHP systematically integrates outcomes of individual filter methods to select the most informative genes for microarray classification. Five individual ranking methods including t-test, entropy, receiver operating characteristic (ROC) curve, Wilcoxon and signal to noise ratio are employed to rank genes. These ranked genes are then considered as inputs for the modified AHP. Additionally, a method that uses fuzzy standard additive model (FSAM) for cancer classification based on genes selected by AHP is also proposed in this paper. Traditional FSAM learning is a hybrid process comprising unsupervised structure learning and supervised parameter tuning. Genetic algorithm (GA) is incorporated in-between unsupervised and supervised training to optimize the number of fuzzy rules. The integration of GA enables FSAM to deal with the high-dimensional-low-sample nature of microarray data and thus enhance the efficiency of the classification. Experiments are carried out on numerous microarray datasets. Results demonstrate the performance dominance of the AHP-based gene selection against the single ranking methods. Furthermore, the combination of AHP-FSAM shows a great accuracy in microarray data classification compared to various competing classifiers. The proposed approach therefore is useful for medical practitioners and clinicians as a decision support system that can be implemented in the real medical practice.

A Hybrid Supervised/Unsupervised Machine Learning Approach to Solar Flare Prediction

NASA Astrophysics Data System (ADS)

Benvenuto, Federico; Piana, Michele; Campi, Cristina; Massone, Anna Maria

2018-01-01

This paper introduces a novel method for flare forecasting, combining prediction accuracy with the ability to identify the most relevant predictive variables. This result is obtained by means of a two-step approach: first, a supervised regularization method for regression, namely, LASSO is applied, where a sparsity-enhancing penalty term allows the identification of the significance with which each data feature contributes to the prediction; then, an unsupervised fuzzy clustering technique for classification, namely, Fuzzy C-Means, is applied, where the regression outcome is partitioned through the minimization of a cost function and without focusing on the optimization of a specific skill score. This approach is therefore hybrid, since it combines supervised and unsupervised learning; realizes classification in an automatic, skill-score-independent way; and provides effective prediction performances even in the case of imbalanced data sets. Its prediction power is verified against NOAA Space Weather Prediction Center data, using as a test set, data in the range between 1996 August and 2010 December and as training set, data in the range between 1988 December and 1996 June. To validate the method, we computed several skill scores typically utilized in flare prediction and compared the values provided by the hybrid approach with the ones provided by several standard (non-hybrid) machine learning methods. The results showed that the hybrid approach performs classification better than all other supervised methods and with an effectiveness comparable to the one of clustering methods; but, in addition, it provides a reliable ranking of the weights with which the data properties contribute to the forecast.
Unsupervised automated high throughput phenotyping of RNAi time-lapse movies.

PubMed

Failmezger, Henrik; Fröhlich, Holger; Tresch, Achim

2013-10-04

Gene perturbation experiments in combination with fluorescence time-lapse cell imaging are a powerful tool in reverse genetics. High content applications require tools for the automated processing of the large amounts of data. These tools include in general several image processing steps, the extraction of morphological descriptors, and the grouping of cells into phenotype classes according to their descriptors. This phenotyping can be applied in a supervised or an unsupervised manner. Unsupervised methods are suitable for the discovery of formerly unknown phenotypes, which are expected to occur in high-throughput RNAi time-lapse screens. We developed an unsupervised phenotyping approach based on Hidden Markov Models (HMMs) with multivariate Gaussian emissions for the detection of knockdown-specific phenotypes in RNAi time-lapse movies. The automated detection of abnormal cell morphologies allows us to assign a phenotypic fingerprint to each gene knockdown. By applying our method to the Mitocheck database, we show that a phenotypic fingerprint is indicative of a gene's function. Our fully unsupervised HMM-based phenotyping is able to automatically identify cell morphologies that are specific for a certain knockdown. Beyond the identification of genes whose knockdown affects cell morphology, phenotypic fingerprints can be used to find modules of functionally related genes.
Infrared vehicle recognition using unsupervised feature learning based on K-feature

NASA Astrophysics Data System (ADS)

Lin, Jin; Tan, Yihua; Xia, Haijiao; Tian, Jinwen

2018-02-01

Subject to the complex battlefield environment, it is difficult to establish a complete knowledge base in practical application of vehicle recognition algorithms. The infrared vehicle recognition is always difficult and challenging, which plays an important role in remote sensing. In this paper we propose a new unsupervised feature learning method based on K-feature to recognize vehicle in infrared images. First, we use the target detection algorithm which is based on the saliency to detect the initial image. Then, the unsupervised feature learning based on K-feature, which is generated by Kmeans clustering algorithm that extracted features by learning a visual dictionary from a large number of samples without label, is calculated to suppress the false alarm and improve the accuracy. Finally, the vehicle target recognition image is finished by some post-processing. Large numbers of experiments demonstrate that the proposed method has satisfy recognition effectiveness and robustness for vehicle recognition in infrared images under complex backgrounds, and it also improve the reliability of it.
Learning viewpoint invariant object representations using a temporal coherence principle.

PubMed

Einhäuser, Wolfgang; Hipp, Jörg; Eggert, Julian; Körner, Edgar; König, Peter

2005-07-01

Invariant object recognition is arguably one of the major challenges for contemporary machine vision systems. In contrast, the mammalian visual system performs this task virtually effortlessly. How can we exploit our knowledge on the biological system to improve artificial systems? Our understanding of the mammalian early visual system has been augmented by the discovery that general coding principles could explain many aspects of neuronal response properties. How can such schemes be transferred to system level performance? In the present study we train cells on a particular variant of the general principle of temporal coherence, the "stability" objective. These cells are trained on unlabeled real-world images without a teaching signal. We show that after training, the cells form a representation that is largely independent of the viewpoint from which the stimulus is looked at. This finding includes generalization to previously unseen viewpoints. The achieved representation is better suited for view-point invariant object classification than the cells' input patterns. This property to facilitate view-point invariant classification is maintained even if training and classification take place in the presence of an--also unlabeled--distractor object. In summary, here we show that unsupervised learning using a general coding principle facilitates the classification of real-world objects, that are not segmented from the background and undergo complex, non-isomorphic, transformations.
Segmentation of fluorescence microscopy cell images using unsupervised mining.

PubMed

Du, Xian; Dua, Sumeet

2010-05-28

The accurate measurement of cell and nuclei contours are critical for the sensitive and specific detection of changes in normal cells in several medical informatics disciplines. Within microscopy, this task is facilitated using fluorescence cell stains, and segmentation is often the first step in such approaches. Due to the complex nature of cell issues and problems inherent to microscopy, unsupervised mining approaches of clustering can be incorporated in the segmentation of cells. In this study, we have developed and evaluated the performance of multiple unsupervised data mining techniques in cell image segmentation. We adapt four distinctive, yet complementary, methods for unsupervised learning, including those based on k-means clustering, EM, Otsu's threshold, and GMAC. Validation measures are defined, and the performance of the techniques is evaluated both quantitatively and qualitatively using synthetic and recently published real data. Experimental results demonstrate that k-means, Otsu's threshold, and GMAC perform similarly, and have more precise segmentation results than EM. We report that EM has higher recall values and lower precision results from under-segmentation due to its Gaussian model assumption. We also demonstrate that these methods need spatial information to segment complex real cell images with a high degree of efficacy, as expected in many medical informatics applications.
Automated classification of dolphin echolocation click types from the Gulf of Mexico.

PubMed

Frasier, Kaitlin E; Roch, Marie A; Soldevilla, Melissa S; Wiggins, Sean M; Garrison, Lance P; Hildebrand, John A

2017-12-01

Delphinids produce large numbers of short duration, broadband echolocation clicks which may be useful for species classification in passive acoustic monitoring efforts. A challenge in echolocation click classification is to overcome the many sources of variability to recognize underlying patterns across many detections. An automated unsupervised network-based classification method was developed to simulate the approach a human analyst uses when categorizing click types: Clusters of similar clicks were identified by incorporating multiple click characteristics (spectral shape and inter-click interval distributions) to distinguish within-type from between-type variation, and identify distinct, persistent click types. Once click types were established, an algorithm for classifying novel detections using existing clusters was tested. The automated classification method was applied to a dataset of 52 million clicks detected across five monitoring sites over two years in the Gulf of Mexico (GOM). Seven distinct click types were identified, one of which is known to be associated with an acoustically identifiable delphinid (Risso's dolphin) and six of which are not yet identified. All types occurred at multiple monitoring locations, but the relative occurrence of types varied, particularly between continental shelf and slope locations. Automatically-identified click types from autonomous seafloor recorders without verifiable species identification were compared with clicks detected on sea-surface towed hydrophone arrays in the presence of visually identified delphinid species. These comparisons suggest potential species identities for the animals producing some echolocation click types. The network-based classification method presented here is effective for rapid, unsupervised delphinid click classification across large datasets in which the click types may not be known a priori.
Automated classification of dolphin echolocation click types from the Gulf of Mexico

PubMed Central

Roch, Marie A.; Soldevilla, Melissa S.; Wiggins, Sean M.; Garrison, Lance P.; Hildebrand, John A.

2017-01-01

Delphinids produce large numbers of short duration, broadband echolocation clicks which may be useful for species classification in passive acoustic monitoring efforts. A challenge in echolocation click classification is to overcome the many sources of variability to recognize underlying patterns across many detections. An automated unsupervised network-based classification method was developed to simulate the approach a human analyst uses when categorizing click types: Clusters of similar clicks were identified by incorporating multiple click characteristics (spectral shape and inter-click interval distributions) to distinguish within-type from between-type variation, and identify distinct, persistent click types. Once click types were established, an algorithm for classifying novel detections using existing clusters was tested. The automated classification method was applied to a dataset of 52 million clicks detected across five monitoring sites over two years in the Gulf of Mexico (GOM). Seven distinct click types were identified, one of which is known to be associated with an acoustically identifiable delphinid (Risso’s dolphin) and six of which are not yet identified. All types occurred at multiple monitoring locations, but the relative occurrence of types varied, particularly between continental shelf and slope locations. Automatically-identified click types from autonomous seafloor recorders without verifiable species identification were compared with clicks detected on sea-surface towed hydrophone arrays in the presence of visually identified delphinid species. These comparisons suggest potential species identities for the animals producing some echolocation click types. The network-based classification method presented here is effective for rapid, unsupervised delphinid click classification across large datasets in which the click types may not be known a priori. PMID:29216184
Unsupervised classification of Space Acceleration Measurement System (SAMS) data using ART2-A

NASA Technical Reports Server (NTRS)

Smith, A. D.; Sinha, A.

1999-01-01

The Space Acceleration Measurement System (SAMS) has been developed by NASA to monitor the microgravity acceleration environment aboard the space shuttle. The amount of data collected by a SAMS unit during a shuttle mission is in the several gigabytes range. Adaptive Resonance Theory 2-A (ART2-A), an unsupervised neural network, has been used to cluster these data and to develop cause and effect relationships among disturbances and the acceleration environment. Using input patterns formed on the basis of power spectral densities (psd), data collected from two missions, STS-050 and STS-057, have been clustered.
Supervised segmentation of microelectrode recording artifacts using power spectral density.

PubMed

Bakstein, Eduard; Schneider, Jakub; Sieger, Tomas; Novak, Daniel; Wild, Jiri; Jech, Robert

2015-08-01

Appropriate detection of clean signal segments in extracellular microelectrode recordings (MER) is vital for maintaining high signal-to-noise ratio in MER studies. Existing alternatives to manual signal inspection are based on unsupervised change-point detection. We present a method of supervised MER artifact classification, based on power spectral density (PSD) and evaluate its performance on a database of 95 labelled MER signals. The proposed method yielded test-set accuracy of 90%, which was close to the accuracy of annotation (94%). The unsupervised methods achieved accuracy of about 77% on both training and testing data.
Maximum Margin Clustering of Hyperspectral Data

NASA Astrophysics Data System (ADS)

Niazmardi, S.; Safari, A.; Homayouni, S.

2013-09-01

In recent decades, large margin methods such as Support Vector Machines (SVMs) are supposed to be the state-of-the-art of supervised learning methods for classification of hyperspectral data. However, the results of these algorithms mainly depend on the quality and quantity of available training data. To tackle down the problems associated with the training data, the researcher put effort into extending the capability of large margin algorithms for unsupervised learning. One of the recent proposed algorithms is Maximum Margin Clustering (MMC). The MMC is an unsupervised SVMs algorithm that simultaneously estimates both the labels and the hyperplane parameters. Nevertheless, the optimization of the MMC algorithm is a non-convex problem. Most of the existing MMC methods rely on the reformulating and the relaxing of the non-convex optimization problem as semi-definite programs (SDP), which are computationally very expensive and only can handle small data sets. Moreover, most of these algorithms are two-class classification, which cannot be used for classification of remotely sensed data. In this paper, a new MMC algorithm is used that solve the original non-convex problem using Alternative Optimization method. This algorithm is also extended for multi-class classification and its performance is evaluated. The results of the proposed algorithm show that the algorithm has acceptable results for hyperspectral data clustering.
Is Nigeria losing its natural vegetation and landscape? Assessing the landuse-landcover change trajectories and effects in Onitsha using remote sensing and GIS

NASA Astrophysics Data System (ADS)

Nwaogu, Chukwudi; Okeke, Onyedikachi J.; Fadipe, Olusola O.; Bashiru, Kehinde A.; Pechanec, Vilém

2017-12-01

Onitsha is one of the largest commercial cities in Africa with its population growth rate increasing arithmetically for the past two decades. This situation has direct and indirect effects on the natural resources including vegetation and water. The study aimed at assessing land use-land cover (LULC) change and its effects on the vegetation and landscape from 1987 to 2015 using geoinformatics. Supervised and unsupervised classifications including maximum likelihood algorithm were performed using ENVI 4.7 and ArcGIS 10.1 versions. The LULC was classified into 7 classes: built-up areas (settlement), waterbody, thick vegetation, light vegetation, riparian vegetation, sand deposit (bare soil) and floodplain. The result revealed that all the three vegetation types decreased in areas throughout the study period while, settlement, sand deposit and floodplain areas have remarkable increase of about 100% in 2015 when compared with the total in 1987. Number of dominant plant species decreased continuously during the study. The overall classification accuracies in 1987, 2002 and 2015 was 90.7%, 92.9% and 95.5% respectively. The overall kappa coefficient of the image classification for 1987, 2002 and 2015 was 0.98, 0.93 and 0.96 respectively. In general, the average classification was above 90%, a proof that the classification was reliable and acceptable.
Nonlinear Deep Kernel Learning for Image Annotation.

PubMed

Jiu, Mingyuan; Sahbi, Hichem

2017-02-08

Multiple kernel learning (MKL) is a widely used technique for kernel design. Its principle consists in learning, for a given support vector classifier, the most suitable convex (or sparse) linear combination of standard elementary kernels. However, these combinations are shallow and often powerless to capture the actual similarity between highly semantic data, especially for challenging classification tasks such as image annotation. In this paper, we redefine multiple kernels using deep multi-layer networks. In this new contribution, a deep multiple kernel is recursively defined as a multi-layered combination of nonlinear activation functions, each one involves a combination of several elementary or intermediate kernels, and results into a positive semi-definite deep kernel. We propose four different frameworks in order to learn the weights of these networks: supervised, unsupervised, kernel-based semisupervised and Laplacian-based semi-supervised. When plugged into support vector machines (SVMs), the resulting deep kernel networks show clear gain, compared to several shallow kernels for the task of image annotation. Extensive experiments and analysis on the challenging ImageCLEF photo annotation benchmark, the COREL5k database and the Banana dataset validate the effectiveness of the proposed method.
Comparison Between Supervised and Unsupervised Classifications of Neuronal Cell Types: A Case Study

PubMed Central

Guerra, Luis; McGarry, Laura M; Robles, Víctor; Bielza, Concha; Larrañaga, Pedro; Yuste, Rafael

2011-01-01

In the study of neural circuits, it becomes essential to discern the different neuronal cell types that build the circuit. Traditionally, neuronal cell types have been classified using qualitative descriptors. More recently, several attempts have been made to classify neurons quantitatively, using unsupervised clustering methods. While useful, these algorithms do not take advantage of previous information known to the investigator, which could improve the classification task. For neocortical GABAergic interneurons, the problem to discern among different cell types is particularly difficult and better methods are needed to perform objective classifications. Here we explore the use of supervised classification algorithms to classify neurons based on their morphological features, using a database of 128 pyramidal cells and 199 interneurons from mouse neocortex. To evaluate the performance of different algorithms we used, as a “benchmark,” the test to automatically distinguish between pyramidal cells and interneurons, defining “ground truth” by the presence or absence of an apical dendrite. We compared hierarchical clustering with a battery of different supervised classification algorithms, finding that supervised classifications outperformed hierarchical clustering. In addition, the selection of subsets of distinguishing features enhanced the classification accuracy for both sets of algorithms. The analysis of selected variables indicates that dendritic features were most useful to distinguish pyramidal cells from interneurons when compared with somatic and axonal morphological variables. We conclude that supervised classification algorithms are better matched to the general problem of distinguishing neuronal cell types when some information on these cell groups, in our case being pyramidal or interneuron, is known a priori. As a spin-off of this methodological study, we provide several methods to automatically distinguish neocortical pyramidal cells from interneurons, based on their morphologies. © 2010 Wiley Periodicals, Inc. Develop Neurobiol 71: 71–82, 2011 PMID:21154911
Automated extraction and classification of time-frequency contours in humpback vocalizations.

PubMed

Ou, Hui; Au, Whitlow W L; Zurk, Lisa M; Lammers, Marc O

2013-01-01

A time-frequency contour extraction and classification algorithm was created to analyze humpback whale vocalizations. The algorithm automatically extracted contours of whale vocalization units by searching for gray-level discontinuities in the spectrogram images. The unit-to-unit similarity was quantified by cross-correlating the contour lines. A library of distinctive humpback units was then generated by applying an unsupervised, cluster-based learning algorithm. The purpose of this study was to provide a fast and automated feature selection tool to describe the vocal signatures of animal groups. This approach could benefit a variety of applications such as species description, identification, and evolution of song structures. The algorithm was tested on humpback whale song data recorded at various locations in Hawaii from 2002 to 2003. Results presented in this paper showed low probability of false alarm (0%-4%) under noisy environments with small boat vessels and snapping shrimp. The classification algorithm was tested on a controlled set of 30 units forming six unit types, and all the units were correctly classified. In a case study on humpback data collected in the Auau Chanel, Hawaii, in 2002, the algorithm extracted 951 units, which were classified into 12 distinctive types.
Texture analysis with statistical methods for wheat ear extraction

NASA Astrophysics Data System (ADS)

Bakhouche, M.; Cointault, F.; Gouton, P.

2007-01-01

In agronomic domain, the simplification of crop counting, necessary for yield prediction and agronomic studies, is an important project for technical institutes such as Arvalis. Although the main objective of our global project is to conceive a mobile robot for natural image acquisition directly in a field, Arvalis has proposed us first to detect by image processing the number of wheat ears in images before to count them, which will allow to obtain the first component of the yield. In this paper we compare different texture image segmentation techniques based on feature extraction by first and higher order statistical methods which have been applied on our images. The extracted features are used for unsupervised pixel classification to obtain the different classes in the image. So, the K-means algorithm is implemented before the choice of a threshold to highlight the ears. Three methods have been tested in this feasibility study with very average error of 6%. Although the evaluation of the quality of the detection is visually done, automatic evaluation algorithms are currently implementing. Moreover, other statistical methods of higher order will be implemented in the future jointly with methods based on spatio-frequential transforms and specific filtering.
Comparing automated classification and digitization approaches to detect change in eelgrass bed extent during restoration of a large river delta

USGS Publications Warehouse

Davenport, Anna Elizabeth; Davis, Jerry D.; Woo, Isa; Grossman, Eric; Barham, Jesse B.; Ellings, Christopher S.; Takekawa, John Y.

2017-01-01

Native eelgrass (Zostera marina) is an important contributor to ecosystem services that supplies cover for juvenile fish, supports a variety of invertebrate prey resources for fish and waterbirds, provides substrate for herring roe consumed by numerous fish and birds, helps stabilize sediment, and sequesters organic carbon. Seagrasses are in decline globally, and monitoring changes in their growth and extent is increasingly valuable to determine impacts from large-scale estuarine restoration and inform blue carbon mapping initiatives. Thus, we examined the efficacy of two remote sensing mapping methods with high-resolution (0.5 m pixel size) color near infrared imagery with ground validation to assess change following major tidal marsh restoration. Automated classification of false color aerial imagery and digitized polygons documented a slight decline in eelgrass area directly after restoration followed by an increase two years later. Classification of sparse and low to medium density eelgrass was confounded in areas with algal cover, however large dense patches of eelgrass were well delineated. Automated classification of aerial imagery from unsupervised and supervised methods provided reasonable accuracies of 73% and hand-digitizing polygons from the same imagery yielded similar results. Visual clues for hand digitizing from the high-resolution imagery provided as reliable a map of dense eelgrass extent as automated image classification. We found that automated classification had no advantages over manual digitization particularly because of the limitations of detecting eelgrass with only three bands of imagery and near infrared.
An Improved Unsupervised Image Segmentation Evaluation Approach Based on - and Over-Segmentation Aware

NASA Astrophysics Data System (ADS)

Su, Tengfei

2018-04-01

In this paper, an unsupervised evaluation scheme for remote sensing image segmentation is developed. Based on a method called under- and over-segmentation aware (UOA), the new approach is improved by overcoming the defect in the part of estimating over-segmentation error. Two cases of such error-prone defect are listed, and edge strength is employed to devise a solution to this issue. Two subsets of high resolution remote sensing images were used to test the proposed algorithm, and the experimental results indicate its superior performance, which is attributed to its improved OSE detection model.
Segmentation of magnetic resonance images using fuzzy algorithms for learning vector quantization.

PubMed

Karayiannis, N B; Pai, P I

1999-02-01

This paper evaluates a segmentation technique for magnetic resonance (MR) images of the brain based on fuzzy algorithms for learning vector quantization (FALVQ). These algorithms perform vector quantization by updating all prototypes of a competitive network through an unsupervised learning process. Segmentation of MR images is formulated as an unsupervised vector quantization process, where the local values of different relaxation parameters form the feature vectors which are represented by a relatively small set of prototypes. The experiments evaluate a variety of FALVQ algorithms in terms of their ability to identify different tissues and discriminate between normal tissues and abnormalities.
A comparative evaluation of supervised and unsupervised representation learning approaches for anaplastic medulloblastoma differentiation

NASA Astrophysics Data System (ADS)

Cruz-Roa, Angel; Arevalo, John; Basavanhally, Ajay; Madabhushi, Anant; González, Fabio

2015-01-01

Learning data representations directly from the data itself is an approach that has shown great success in different pattern recognition problems, outperforming state-of-the-art feature extraction schemes for different tasks in computer vision, speech recognition and natural language processing. Representation learning applies unsupervised and supervised machine learning methods to large amounts of data to find building-blocks that better represent the information in it. Digitized histopathology images represents a very good testbed for representation learning since it involves large amounts of high complex, visual data. This paper presents a comparative evaluation of different supervised and unsupervised representation learning architectures to specifically address open questions on what type of learning architectures (deep or shallow), type of learning (unsupervised or supervised) is optimal. In this paper we limit ourselves to addressing these questions in the context of distinguishing between anaplastic and non-anaplastic medulloblastomas from routine haematoxylin and eosin stained images. The unsupervised approaches evaluated were sparse autoencoders and topographic reconstruct independent component analysis, and the supervised approach was convolutional neural networks. Experimental results show that shallow architectures with more neurons are better than deeper architectures without taking into account local space invariances and that topographic constraints provide useful invariant features in scale and rotations for efficient tumor differentiation.
A new local-global approach for classification.

PubMed

Peres, R T; Pedreira, C E

2010-09-01

In this paper, we propose a new local-global pattern classification scheme that combines supervised and unsupervised approaches, taking advantage of both, local and global environments. We understand as global methods the ones concerned with the aim of constructing a model for the whole problem space using the totality of the available observations. Local methods focus into sub regions of the space, possibly using an appropriately selected subset of the sample. In the proposed method, the sample is first divided in local cells by using a Vector Quantization unsupervised algorithm, the LBG (Linde-Buzo-Gray). In a second stage, the generated assemblage of much easier problems is locally solved with a scheme inspired by Bayes' rule. Four classification methods were implemented for comparison purposes with the proposed scheme: Learning Vector Quantization (LVQ); Feedforward Neural Networks; Support Vector Machine (SVM) and k-Nearest Neighbors. These four methods and the proposed scheme were implemented in eleven datasets, two controlled experiments, plus nine public available datasets from the UCI repository. The proposed method has shown a quite competitive performance when compared to these classical and largely used classifiers. Our method is simple concerning understanding and implementation and is based on very intuitive concepts. Copyright 2010 Elsevier Ltd. All rights reserved.

Unsupervised Biomedical Named Entity Recognition: Experiments with Clinical and Biological Texts

PubMed Central

Zhang, Shaodian; Elhadad, Nóemie

2013-01-01

Named entity recognition is a crucial component of biomedical natural language processing, enabling information extraction and ultimately reasoning over and knowledge discovery from text. Much progress has been made in the design of rule-based and supervised tools, but they are often genre and task dependent. As such, adapting them to different genres of text or identifying new types of entities requires major effort in re-annotation or rule development. In this paper, we propose an unsupervised approach to extracting named entities from biomedical text. We describe a stepwise solution to tackle the challenges of entity boundary detection and entity type classification without relying on any handcrafted rules, heuristics, or annotated data. A noun phrase chunker followed by a filter based on inverse document frequency extracts candidate entities from free text. Classification of candidate entities into categories of interest is carried out by leveraging principles from distributional semantics. Experiments show that our system, especially the entity classification step, yields competitive results on two popular biomedical datasets of clinical notes and biological literature, and outperforms a baseline dictionary match approach. Detailed error analysis provides a road map for future work. PMID:23954592
Automated attribution of remotely-sensed ecological disturbances using spatial and temporal characteristics of common disturbance classes.

NASA Astrophysics Data System (ADS)

Cooper, L. A.; Ballantyne, A.

2017-12-01

Forest disturbances are critical components of ecosystems. Knowledge of their prevalence and impacts is necessary to accurately describe forest health and ecosystem services through time. While there are currently several methods available to identify and describe forest disturbances, especially those which occur in North America, the process remains inefficient and inaccessible in many parts of the world. Here, we introduce a preliminary approach to streamline and automate both the detection and attribution of forest disturbances. We use a combination of the Breaks for Additive Season and Trend (BFAST) detection algorithm to detect disturbances in combination with supervised and unsupervised classification algorithms to attribute the detections to disturbance classes. Both spatial and temporal disturbance characteristics are derived and utilized for the goal of automating the disturbance attribution process. The resulting preliminary algorithm is applied to up-scaled (100m) Landsat data for several different ecosystems in North America, with varying success. Our results indicate that supervised classification is more reliable than unsupervised classification, but that limited training data are required for a region. Future work will improve the algorithm through refining and validating at sites within North America before applying this approach globally.
A recurrent neural network for classification of unevenly sampled variable stars

NASA Astrophysics Data System (ADS)

Naul, Brett; Bloom, Joshua S.; Pérez, Fernando; van der Walt, Stéfan

2018-02-01

Astronomical surveys of celestial sources produce streams of noisy time series measuring flux versus time (`light curves'). Unlike in many other physical domains, however, large (and source-specific) temporal gaps in data arise naturally due to intranight cadence choices as well as diurnal and seasonal constraints1-5. With nightly observations of millions of variable stars and transients from upcoming surveys4,6, efficient and accurate discovery and classification techniques on noisy, irregularly sampled data must be employed with minimal human-in-the-loop involvement. Machine learning for inference tasks on such data traditionally requires the laborious hand-coding of domain-specific numerical summaries of raw data (`features')7. Here, we present a novel unsupervised autoencoding recurrent neural network8 that makes explicit use of sampling times and known heteroskedastic noise properties. When trained on optical variable star catalogues, this network produces supervised classification models that rival other best-in-class approaches. We find that autoencoded features learned in one time-domain survey perform nearly as well when applied to another survey. These networks can continue to learn from new unlabelled observations and may be used in other unsupervised tasks, such as forecasting and anomaly detection.
Automated egg grading system using computer vision: Investigation on weight measure versus shape parameters

NASA Astrophysics Data System (ADS)

Nasir, Ahmad Fakhri Ab; Suhaila Sabarudin, Siti; Majeed, Anwar P. P. Abdul; Ghani, Ahmad Shahrizan Abdul

2018-04-01

Chicken egg is a source of food of high demand by humans. Human operators cannot work perfectly and continuously when conducting egg grading. Instead of an egg grading system using weight measure, an automatic system for egg grading using computer vision (using egg shape parameter) can be used to improve the productivity of egg grading. However, early hypothesis has indicated that more number of egg classes will change when using egg shape parameter compared with using weight measure. This paper presents the comparison of egg classification by the two above-mentioned methods. Firstly, 120 images of chicken eggs of various grades (A–D) produced in Malaysia are captured. Then, the egg images are processed using image pre-processing techniques, such as image cropping, smoothing and segmentation. Thereafter, eight egg shape features, including area, major axis length, minor axis length, volume, diameter and perimeter, are extracted. Lastly, feature selection (information gain ratio) and feature extraction (principal component analysis) are performed using k-nearest neighbour classifier in the classification process. Two methods, namely, supervised learning (using weight measure as graded by egg supplier) and unsupervised learning (using egg shape parameters as graded by ourselves), are conducted to execute the experiment. Clustering results reveal many changes in egg classes after performing shape-based grading. On average, the best recognition results using shape-based grading label is 94.16% while using weight-based label is 44.17%. As conclusion, automated egg grading system using computer vision is better by implementing shape-based features since it uses image meanwhile the weight parameter is more suitable by using weight grading system.
Unsupervised classification of multivariate geostatistical data: Two algorithms

NASA Astrophysics Data System (ADS)

Romary, Thomas; Ors, Fabien; Rivoirard, Jacques; Deraisme, Jacques

2015-12-01

With the increasing development of remote sensing platforms and the evolution of sampling facilities in mining and oil industry, spatial datasets are becoming increasingly large, inform a growing number of variables and cover wider and wider areas. Therefore, it is often necessary to split the domain of study to account for radically different behaviors of the natural phenomenon over the domain and to simplify the subsequent modeling step. The definition of these areas can be seen as a problem of unsupervised classification, or clustering, where we try to divide the domain into homogeneous domains with respect to the values taken by the variables in hand. The application of classical clustering methods, designed for independent observations, does not ensure the spatial coherence of the resulting classes. Image segmentation methods, based on e.g. Markov random fields, are not adapted to irregularly sampled data. Other existing approaches, based on mixtures of Gaussian random functions estimated via the expectation-maximization algorithm, are limited to reasonable sample sizes and a small number of variables. In this work, we propose two algorithms based on adaptations of classical algorithms to multivariate geostatistical data. Both algorithms are model free and can handle large volumes of multivariate, irregularly spaced data. The first one proceeds by agglomerative hierarchical clustering. The spatial coherence is ensured by a proximity condition imposed for two clusters to merge. This proximity condition relies on a graph organizing the data in the coordinates space. The hierarchical algorithm can then be seen as a graph-partitioning algorithm. Following this interpretation, a spatial version of the spectral clustering algorithm is also proposed. The performances of both algorithms are assessed on toy examples and a mining dataset.
Incrementally learning objects by touch: online discriminative and generative models for tactile-based recognition.

PubMed

Soh, Harold; Demiris, Yiannis

2014-01-01

Human beings not only possess the remarkable ability to distinguish objects through tactile feedback but are further able to improve upon recognition competence through experience. In this work, we explore tactile-based object recognition with learners capable of incremental learning. Using the sparse online infinite Echo-State Gaussian process (OIESGP), we propose and compare two novel discriminative and generative tactile learners that produce probability distributions over objects during object grasping/palpation. To enable iterative improvement, our online methods incorporate training samples as they become available. We also describe incremental unsupervised learning mechanisms, based on novelty scores and extreme value theory, when teacher labels are not available. We present experimental results for both supervised and unsupervised learning tasks using the iCub humanoid, with tactile sensors on its five-fingered anthropomorphic hand, and 10 different object classes. Our classifiers perform comparably to state-of-the-art methods (C4.5 and SVM classifiers) and findings indicate that tactile signals are highly relevant for making accurate object classifications. We also show that accurate "early" classifications are possible using only 20-30 percent of the grasp sequence. For unsupervised learning, our methods generate high quality clusterings relative to the widely-used sequential k-means and self-organising map (SOM), and we present analyses into the differences between the approaches.
Learning relevant features of data with multi-scale tensor networks

NASA Astrophysics Data System (ADS)

Miles Stoudenmire, E.

2018-07-01

Inspired by coarse-graining approaches used in physics, we show how similar algorithms can be adapted for data. The resulting algorithms are based on layered tree tensor networks and scale linearly with both the dimension of the input and the training set size. Computing most of the layers with an unsupervised algorithm, then optimizing just the top layer for supervised classification of the MNIST and fashion MNIST data sets gives very good results. We also discuss mixing a prior guess for supervised weights together with an unsupervised representation of the data, yielding a smaller number of features nevertheless able to give good performance.
Watermarking techniques for electronic delivery of remote sensing images

NASA Astrophysics Data System (ADS)

Barni, Mauro; Bartolini, Franco; Magli, Enrico; Olmo, Gabriella

2002-09-01

Earth observation missions have recently attracted a growing interest, mainly due to the large number of possible applications capable of exploiting remotely sensed data and images. Along with the increase of market potential, the need arises for the protection of the image products. Such a need is a very crucial one, because the Internet and other public/private networks have become preferred means of data exchange. A critical issue arising when dealing with digital image distribution is copyright protection. Such a problem has been largely addressed by resorting to watermarking technology. A question that obviously arises is whether the requirements imposed by remote sensing imagery are compatible with existing watermarking techniques. On the basis of these motivations, the contribution of this work is twofold: assessment of the requirements imposed by remote sensing applications on watermark-based copyright protection, and modification of two well-established digital watermarking techniques to meet such constraints. More specifically, the concept of near-lossless watermarking is introduced and two possible algorithms matching such a requirement are presented. Experimental results are shown to measure the impact of watermark introduction on a typical remote sensing application, i.e., unsupervised image classification.
Hyperspectral Imaging and K-Means Classification for Histologic Evaluation of Ductal Carcinoma In Situ.

PubMed

Khouj, Yasser; Dawson, Jeremy; Coad, James; Vona-Davis, Linda

2018-01-01

Hyperspectral imaging (HSI) is a non-invasive optical imaging modality that shows the potential to aid pathologists in breast cancer diagnoses cases. In this study, breast cancer tissues from different patients were imaged by a hyperspectral system to detect spectral differences between normal and breast cancer tissues. Tissue samples mounted on slides were identified from 10 different patients. Samples from each patient included both normal and ductal carcinoma tissue, both stained with hematoxylin and eosin stain and unstained. Slides were imaged using a snapshot HSI system, and the spectral reflectance differences were evaluated. Analysis of the spectral reflectance values indicated that wavelengths near 550 nm showed the best differentiation between tissue types. This information was used to train image processing algorithms using supervised and unsupervised data. The K-means method was applied to the hyperspectral data cubes, and successfully detected spectral tissue differences with sensitivity of 85.45%, and specificity of 94.64% with true negative rate of 95.8%, and false positive rate of 4.2%. These results were verified by ground-truth marking of the tissue samples by a pathologist. In the hyperspectral image analysis, the image processing algorithm, K-means, shows the greatest potential for building a semi-automated system that could identify and sort between normal and ductal carcinoma in situ tissues.
Evaluating Mixture Modeling for Clustering: Recommendations and Cautions

ERIC Educational Resources Information Center

Steinley, Douglas; Brusco, Michael J.

2011-01-01

This article provides a large-scale investigation into several of the properties of mixture-model clustering techniques (also referred to as latent class cluster analysis, latent profile analysis, model-based clustering, probabilistic clustering, Bayesian classification, unsupervised learning, and finite mixture models; see Vermunt & Magdison,…
Remote Sensing Monitoring of Changes in Soil Salinity: A Case Study in Inner Mongolia, China.

PubMed

Wu, Jingwei; Vincent, Bernard; Yang, Jinzhong; Bouarfa, Sami; Vidal, Alain

2008-11-07

This study used archived remote sensing images to depict the history of changes in soil salinity in the Hetao Irrigation District in Inner Mongolia, China, with the purpose of linking these changes with land and water management practices and to draw lessons for salinity control. Most data came from LANDSAT satellite images taken in 1973, 1977, 1988, 1991, 1996, 2001, and 2006. In these years salt-affected areas were detected using a normal supervised classification method. Corresponding cropped areas were detected from NVDI (Normalized Difference Vegetation Index) values using an unsupervised method. Field samples and agricultural statistics were used to estimate the accuracy of the classification. Historical data concerning irrigation/drainage and the groundwater table were used to analyze the relation between changes in soil salinity and land and water management practices. Results showed that: (1) the overall accuracy of remote sensing in detecting soil salinity was 90.2%, and in detecting cropped area, 98%; (2) the installation/innovation of the drainage system did help to control salinity; and (3) a low ratio of cropped land helped control salinity in the Hetao Irrigation District. These findings suggest that remote sensing is a useful tool to detect soil salinity and has potential in evaluating and improving land and water management practices.
Hierarchical Adaptive Means (HAM) clustering for hardware-efficient, unsupervised and real-time spike sorting.

PubMed

Paraskevopoulou, Sivylla E; Wu, Di; Eftekhar, Amir; Constandinou, Timothy G

2014-09-30

This work presents a novel unsupervised algorithm for real-time adaptive clustering of neural spike data (spike sorting). The proposed Hierarchical Adaptive Means (HAM) clustering method combines centroid-based clustering with hierarchical cluster connectivity to classify incoming spikes using groups of clusters. It is described how the proposed method can adaptively track the incoming spike data without requiring any past history, iteration or training and autonomously determines the number of spike classes. Its performance (classification accuracy) has been tested using multiple datasets (both simulated and recorded) achieving a near-identical accuracy compared to k-means (using 10-iterations and provided with the number of spike classes). Also, its robustness in applying to different feature extraction methods has been demonstrated by achieving classification accuracies above 80% across multiple datasets. Last but crucially, its low complexity, that has been quantified through both memory and computation requirements makes this method hugely attractive for future hardware implementation. Copyright © 2014 Elsevier B.V. All rights reserved.
Conduction Delay Learning Model for Unsupervised and Supervised Classification of Spatio-Temporal Spike Patterns

PubMed Central

Matsubara, Takashi

2017-01-01

Precise spike timing is considered to play a fundamental role in communications and signal processing in biological neural networks. Understanding the mechanism of spike timing adjustment would deepen our understanding of biological systems and enable advanced engineering applications such as efficient computational architectures. However, the biological mechanisms that adjust and maintain spike timing remain unclear. Existing algorithms adopt a supervised approach, which adjusts the axonal conduction delay and synaptic efficacy until the spike timings approximate the desired timings. This study proposes a spike timing-dependent learning model that adjusts the axonal conduction delay and synaptic efficacy in both unsupervised and supervised manners. The proposed learning algorithm approximates the Expectation-Maximization algorithm, and classifies the input data encoded into spatio-temporal spike patterns. Even in the supervised classification, the algorithm requires no external spikes indicating the desired spike timings unlike existing algorithms. Furthermore, because the algorithm is consistent with biological models and hypotheses found in existing biological studies, it could capture the mechanism underlying biological delay learning. PMID:29209191
Conduction Delay Learning Model for Unsupervised and Supervised Classification of Spatio-Temporal Spike Patterns.

PubMed

Matsubara, Takashi

2017-01-01

Precise spike timing is considered to play a fundamental role in communications and signal processing in biological neural networks. Understanding the mechanism of spike timing adjustment would deepen our understanding of biological systems and enable advanced engineering applications such as efficient computational architectures. However, the biological mechanisms that adjust and maintain spike timing remain unclear. Existing algorithms adopt a supervised approach, which adjusts the axonal conduction delay and synaptic efficacy until the spike timings approximate the desired timings. This study proposes a spike timing-dependent learning model that adjusts the axonal conduction delay and synaptic efficacy in both unsupervised and supervised manners. The proposed learning algorithm approximates the Expectation-Maximization algorithm, and classifies the input data encoded into spatio-temporal spike patterns. Even in the supervised classification, the algorithm requires no external spikes indicating the desired spike timings unlike existing algorithms. Furthermore, because the algorithm is consistent with biological models and hypotheses found in existing biological studies, it could capture the mechanism underlying biological delay learning.
SAR image segmentation using skeleton-based fuzzy clustering

NASA Astrophysics Data System (ADS)

Cao, Yun Yi; Chen, Yan Qiu

2003-06-01

SAR image segmentation can be converted to a clustering problem in which pixels or small patches are grouped together based on local feature information. In this paper, we present a novel framework for segmentation. The segmentation goal is achieved by unsupervised clustering upon characteristic descriptors extracted from local patches. The mixture model of characteristic descriptor, which combines intensity and texture feature, is investigated. The unsupervised algorithm is derived from the recently proposed Skeleton-Based Data Labeling method. Skeletons are constructed as prototypes of clusters to represent arbitrary latent structures in image data. Segmentation using Skeleton-Based Fuzzy Clustering is able to detect the types of surfaces appeared in SAR images automatically without any user input.
A Comparative Evaluation of Unsupervised Anomaly Detection Algorithms for Multivariate Data.

PubMed

Goldstein, Markus; Uchida, Seiichi

2016-01-01

Anomaly detection is the process of identifying unexpected items or events in datasets, which differ from the norm. In contrast to standard classification tasks, anomaly detection is often applied on unlabeled data, taking only the internal structure of the dataset into account. This challenge is known as unsupervised anomaly detection and is addressed in many practical applications, for example in network intrusion detection, fraud detection as well as in the life science and medical domain. Dozens of algorithms have been proposed in this area, but unfortunately the research community still lacks a comparative universal evaluation as well as common publicly available datasets. These shortcomings are addressed in this study, where 19 different unsupervised anomaly detection algorithms are evaluated on 10 different datasets from multiple application domains. By publishing the source code and the datasets, this paper aims to be a new well-funded basis for unsupervised anomaly detection research. Additionally, this evaluation reveals the strengths and weaknesses of the different approaches for the first time. Besides the anomaly detection performance, computational effort, the impact of parameter settings as well as the global/local anomaly detection behavior is outlined. As a conclusion, we give an advise on algorithm selection for typical real-world tasks.
Knowledge-Based Topic Model for Unsupervised Object Discovery and Localization.

PubMed

Niu, Zhenxing; Hua, Gang; Wang, Le; Gao, Xinbo

Unsupervised object discovery and localization is to discover some dominant object classes and localize all of object instances from a given image collection without any supervision. Previous work has attempted to tackle this problem with vanilla topic models, such as latent Dirichlet allocation (LDA). However, in those methods no prior knowledge for the given image collection is exploited to facilitate object discovery. On the other hand, the topic models used in those methods suffer from the topic coherence issue-some inferred topics do not have clear meaning, which limits the final performance of object discovery. In this paper, prior knowledge in terms of the so-called must-links are exploited from Web images on the Internet. Furthermore, a novel knowledge-based topic model, called LDA with mixture of Dirichlet trees, is proposed to incorporate the must-links into topic modeling for object discovery. In particular, to better deal with the polysemy phenomenon of visual words, the must-link is re-defined as that one must-link only constrains one or some topic(s) instead of all topics, which leads to significantly improved topic coherence. Moreover, the must-links are built and grouped with respect to specific object classes, thus the must-links in our approach are semantic-specific , which allows to more efficiently exploit discriminative prior knowledge from Web images. Extensive experiments validated the efficiency of our proposed approach on several data sets. It is shown that our method significantly improves topic coherence and outperforms the unsupervised methods for object discovery and localization. In addition, compared with discriminative methods, the naturally existing object classes in the given image collection can be subtly discovered, which makes our approach well suited for realistic applications of unsupervised object discovery.Unsupervised object discovery and localization is to discover some dominant object classes and localize all of object instances from a given image collection without any supervision. Previous work has attempted to tackle this problem with vanilla topic models, such as latent Dirichlet allocation (LDA). However, in those methods no prior knowledge for the given image collection is exploited to facilitate object discovery. On the other hand, the topic models used in those methods suffer from the topic coherence issue-some inferred topics do not have clear meaning, which limits the final performance of object discovery. In this paper, prior knowledge in terms of the so-called must-links are exploited from Web images on the Internet. Furthermore, a novel knowledge-based topic model, called LDA with mixture of Dirichlet trees, is proposed to incorporate the must-links into topic modeling for object discovery. In particular, to better deal with the polysemy phenomenon of visual words, the must-link is re-defined as that one must-link only constrains one or some topic(s) instead of all topics, which leads to significantly improved topic coherence. Moreover, the must-links are built and grouped with respect to specific object classes, thus the must-links in our approach are semantic-specific , which allows to more efficiently exploit discriminative prior knowledge from Web images. Extensive experiments validated the efficiency of our proposed approach on several data sets. It is shown that our method significantly improves topic coherence and outperforms the unsupervised methods for object discovery and localization. In addition, compared with discriminative methods, the naturally existing object classes in the given image collection can be subtly discovered, which makes our approach well suited for realistic applications of unsupervised object discovery.
Color normalization of histology slides using graph regularized sparse NMF

NASA Astrophysics Data System (ADS)

Sha, Lingdao; Schonfeld, Dan; Sethi, Amit

2017-03-01

Computer based automatic medical image processing and quantification are becoming popular in digital pathology. However, preparation of histology slides can vary widely due to differences in staining equipment, procedures and reagents, which can reduce the accuracy of algorithms that analyze their color and texture information. To re- duce the unwanted color variations, various supervised and unsupervised color normalization methods have been proposed. Compared with supervised color normalization methods, unsupervised color normalization methods have advantages of time and cost efficient and universal applicability. Most of the unsupervised color normaliza- tion methods for histology are based on stain separation. Based on the fact that stain concentration cannot be negative and different parts of the tissue absorb different stains, nonnegative matrix factorization (NMF), and particular its sparse version (SNMF), are good candidates for stain separation. However, most of the existing unsupervised color normalization method like PCA, ICA, NMF and SNMF fail to consider important information about sparse manifolds that its pixels occupy, which could potentially result in loss of texture information during color normalization. Manifold learning methods like Graph Laplacian have proven to be very effective in interpreting high-dimensional data. In this paper, we propose a novel unsupervised stain separation method called graph regularized sparse nonnegative matrix factorization (GSNMF). By considering the sparse prior of stain concentration together with manifold information from high-dimensional image data, our method shows better performance in stain color deconvolution than existing unsupervised color deconvolution methods, especially in keeping connected texture information. To utilized the texture information, we construct a nearest neighbor graph between pixels within a spatial area of an image based on their distances using heat kernal in lαβ space. The representation of a pixel in the stain density space is constrained to follow the feature distance of the pixel to pixels in the neighborhood graph. Utilizing color matrix transfer method with the stain concentrations found using our GSNMF method, the color normalization performance was also better than existing methods.
Classification of river water pollution using Hyperion data

NASA Astrophysics Data System (ADS)

Kar, Soumyashree; Rathore, V. S.; Champati ray, P. K.; Sharma, Richa; Swain, S. K.

2016-06-01

A novel attempt is made to use hyperspectral remote sensing to identify the spatial variability of metal pollutants present in river water. It was also attempted to classify the hyperspectral image - Earth Observation-1 (EO-1) Hyperion data of an 8 km stretch of the river Yamuna, near Allahabad city in India depending on its chemical composition. For validating image analysis results, a total of 10 water samples were collected and chemically analyzed using Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES). Two different spectral libraries from field and image data were generated for the 10 sample locations. Advanced per-pixel supervised classifications such as Spectral Angle Mapper (SAM), SAM target finder using BandMax and Support Vector Machine (SVM) were carried out along with the unsupervised clustering procedure - Iterative Self-Organizing Data Analysis Technique (ISODATA). The results were compared and assessed with respect to ground data. Analytical Spectral Devices (ASD), Inc. spectroradiometer, FieldSpec 4 was used to generate the spectra of the water samples which were compiled into a spectral library and used for Spectral Absorption Depth (SAD) analysis. The spectral depth pattern of image and field spectral libraries was found to be highly correlated (correlation coefficient, R2 = 0.99) which validated the image analysis results with respect to the ground data. Further, we carried out a multivariate regression analysis to assess the varying concentrations of metal ions present in water based on the spectral depth of the corresponding absorption feature. Spectral Absorption Depth (SAD) analysis along with metal analysis of field data revealed the order in which the metals affected the river pollution, which was in conformity with the findings of Central Pollution Control Board (CPCB). Therefore, it is concluded that hyperspectral imaging provides opportunity that can be used for satellite based remote monitoring of water quality from space.
Multi-site evaluation of IKONOS data for classification of tropical coral reef environments

USGS Publications Warehouse

Andrefouet, S.; Kramer, Philip; Torres-Pulliza, D.; Joyce, K.E.; Hochberg, E.J.; Garza-Perez, R.; Mumby, P.J.; Riegl, Bernhard; Yamano, H.; White, W.H.; Zubia, M.; Brock, J.C.; Phinn, S.R.; Naseer, A.; Hatcher, B.G.; Muller-Karger, F. E.

2003-01-01

Ten IKONOS images of different coral reef sites distributed around the world were processed to assess the potential of 4-m resolution multispectral data for coral reef habitat mapping. Complexity of reef environments, established by field observation, ranged from 3 to 15 classes of benthic habitats containing various combinations of sediments, carbonate pavement, seagrass, algae, and corals in different geomorphologic zones (forereef, lagoon, patch reef, reef flats). Processing included corrections for sea surface roughness and bathymetry, unsupervised or supervised classification, and accuracy assessment based on ground-truth data. IKONOS classification results were compared with classified Landsat 7 imagery for simple to moderate complexity of reef habitats (5-11 classes). For both sensors, overall accuracies of the classifications show a general linear trend of decreasing accuracy with increasing habitat complexity. The IKONOS sensor performed better, with a 15-20% improvement in accuracy compared to Landsat. For IKONOS, overall accuracy was 77% for 4-5 classes, 71% for 7-8 classes, 65% in 9-11 classes, and 53% for more than 13 classes. The Landsat classification accuracy was systematically lower, with an average of 56% for 5-10 classes. Within this general trend, inter-site comparisons and specificities demonstrate the benefits of different approaches. Pre-segmentation of the different geomorphologic zones and depth correction provided different advantages in different environments. Our results help guide scientists and managers in applying IKONOS-class data for coral reef mapping applications. ?? 2003 Elsevier Inc. All rights reserved.

Employing broadband spectra and cluster analysis to assess thermal defoliation of cotton

USDA-ARS?s Scientific Manuscript database

Growers and field scouts need assistance in surveying cotton (Gossypium hirsutum L.) fields subjected to thermal defoliation to reap the benefits provided by this nonchemical defoliation method. A study was conducted to evaluate broadband spectral data and unsupervised classification as tools for s...
Understanding Student Language: An Unsupervised Dialogue Act Classification Approach

ERIC Educational Resources Information Center

Ezen-Can, Aysu; Boyer, Kristy Elizabeth

2015-01-01

Within the landscape of educational data, textual natural language is an increasingly vast source of learning-centered interactions. In natural language dialogue, student contributions hold important information about knowledge and goals. Automatically modeling the dialogue act of these student utterances is crucial for scaling natural language…
Measuring Category Intuitiveness in Unconstrained Categorization Tasks

ERIC Educational Resources Information Center

Pothos, Emmanuel M.; Perlman, Amotz; Bailey, Todd M.; Kurtz, Ken; Edwards, Darren J.; Hines, Peter; McDonnell, John V.

2011-01-01

What makes a category seem natural or intuitive? In this paper, an unsupervised categorization task was employed to examine observer agreement concerning the categorization of nine different stimulus sets. The stimulus sets were designed to capture different intuitions about classification structure. The main empirical index of category…
Land use and land cover classification for rural residential areas in China using soft-probability cascading of multifeatures

NASA Astrophysics Data System (ADS)

Zhang, Bin; Liu, Yueyan; Zhang, Zuyu; Shen, Yonglin

2017-10-01

A multifeature soft-probability cascading scheme to solve the problem of land use and land cover (LULC) classification using high-spatial-resolution images to map rural residential areas in China is proposed. The proposed method is used to build midlevel LULC features. Local features are frequently considered as low-level feature descriptors in a midlevel feature learning method. However, spectral and textural features, which are very effective low-level features, are neglected. The acquisition of the dictionary of sparse coding is unsupervised, and this phenomenon reduces the discriminative power of the midlevel feature. Thus, we propose to learn supervised features based on sparse coding, a support vector machine (SVM) classifier, and a conditional random field (CRF) model to utilize the different effective low-level features and improve the discriminability of midlevel feature descriptors. First, three kinds of typical low-level features, namely, dense scale-invariant feature transform, gray-level co-occurrence matrix, and spectral features, are extracted separately. Second, combined with sparse coding and the SVM classifier, the probabilities of the different LULC classes are inferred to build supervised feature descriptors. Finally, the CRF model, which consists of two parts: unary potential and pairwise potential, is employed to construct an LULC classification map. Experimental results show that the proposed classification scheme can achieve impressive performance when the total accuracy reached about 87%.
Automated and unsupervised detection of malarial parasites in microscopic images.

PubMed

Purwar, Yashasvi; Shah, Sirish L; Clarke, Gwen; Almugairi, Areej; Muehlenbachs, Atis

2011-12-13

Malaria is a serious infectious disease. According to the World Health Organization, it is responsible for nearly one million deaths each year. There are various techniques to diagnose malaria of which manual microscopy is considered to be the gold standard. However due to the number of steps required in manual assessment, this diagnostic method is time consuming (leading to late diagnosis) and prone to human error (leading to erroneous diagnosis), even in experienced hands. The focus of this study is to develop a robust, unsupervised and sensitive malaria screening technique with low material cost and one that has an advantage over other techniques in that it minimizes human reliance and is, therefore, more consistent in applying diagnostic criteria. A method based on digital image processing of Giemsa-stained thin smear image is developed to facilitate the diagnostic process. The diagnosis procedure is divided into two parts; enumeration and identification. The image-based method presented here is designed to automate the process of enumeration and identification; with the main advantage being its ability to carry out the diagnosis in an unsupervised manner and yet have high sensitivity and thus reducing cases of false negatives. The image based method is tested over more than 500 images from two independent laboratories. The aim is to distinguish between positive and negative cases of malaria using thin smear blood slide images. Due to the unsupervised nature of method it requires minimal human intervention thus speeding up the whole process of diagnosis. Overall sensitivity to capture cases of malaria is 100% and specificity ranges from 50-88% for all species of malaria parasites. Image based screening method will speed up the whole process of diagnosis and is more advantageous over laboratory procedures that are prone to errors and where pathological expertise is minimal. Further this method provides a consistent and robust way of generating the parasite clearance curves.
Neural-network classifiers for automatic real-world aerial image recognition

NASA Astrophysics Data System (ADS)

Greenberg, Shlomo; Guterman, Hugo

1996-08-01

We describe the application of the multilayer perceptron (MLP) network and a version of the adaptive resonance theory version 2-A (ART 2-A) network to the problem of automatic aerial image recognition (AAIR). The classification of aerial images, independent of their positions and orientations, is required for automatic tracking and target recognition. Invariance is achieved by the use of different invariant feature spaces in combination with supervised and unsupervised neural networks. The performance of neural-network-based classifiers in conjunction with several types of invariant AAIR global features, such as the Fourier-transform space, Zernike moments, central moments, and polar transforms, are examined. The advantages of this approach are discussed. The performance of the MLP network is compared with that of a classical correlator. The MLP neural-network correlator outperformed the binary phase-only filter (BPOF) correlator. It was found that the ART 2-A distinguished itself with its speed and its low number of required training vectors. However, only the MLP classifier was able to deal with a combination of shift and rotation geometric distortions.
Neural-network classifiers for automatic real-world aerial image recognition.

PubMed

Greenberg, S; Guterman, H

1996-08-10

We describe the application of the multilayer perceptron (MLP) network and a version of the adaptive resonance theory version 2-A (ART 2-A) network to the problem of automatic aerial image recognition (AAIR). The classification of aerial images, independent of their positions and orientations, is required for automatic tracking and target recognition. Invariance is achieved by the use of different invariant feature spaces in combination with supervised and unsupervised neural networks. The performance of neural-network-based classifiers in conjunction with several types of invariant AAIR global features, such as the Fourier-transform space, Zernike moments, central moments, and polar transforms, are examined. The advantages of this approach are discussed. The performance of the MLP network is compared with that of a classical correlator. The MLP neural-network correlator outperformed the binary phase-only filter (BPOF) correlator. It was found that the ART 2-A distinguished itself with its speed and its low number of required training vectors. However, only the MLP classifier was able to deal with a combination of shift and rotation geometric distortions.
Cloud classification from satellite data using a fuzzy sets algorithm: A polar example

NASA Technical Reports Server (NTRS)

Key, J. R.; Maslanik, J. A.; Barry, R. G.

1988-01-01

Where spatial boundaries between phenomena are diffuse, classification methods which construct mutually exclusive clusters seem inappropriate. The Fuzzy c-means (FCM) algorithm assigns each observation to all clusters, with membership values as a function of distance to the cluster center. The FCM algorithm is applied to AVHRR data for the purpose of classifying polar clouds and surfaces. Careful analysis of the fuzzy sets can provide information on which spectral channels are best suited to the classification of particular features, and can help determine likely areas of misclassification. General agreement in the resulting classes and cloud fraction was found between the FCM algorithm, a manual classification, and an unsupervised maximum likelihood classifier.
Residential roof condition assessment system using deep learning

NASA Astrophysics Data System (ADS)

Wang, Fan; Kerekes, John P.; Xu, Zhuoyi; Wang, Yandong

2018-01-01

The emergence of high resolution (HR) and ultra high resolution (UHR) airborne remote sensing imagery is enabling humans to move beyond traditional land cover analysis applications to the detailed characterization of surface objects. A residential roof condition assessment method using techniques from deep learning is presented. The proposed method operates on individual roofs and divides the task into two stages: (1) roof segmentation, followed by (2) condition classification of the segmented roof regions. As the first step in this process, a self-tuning method is proposed to segment the images into small homogeneous areas. The segmentation is initialized with simple linear iterative clustering followed by deep learned feature extraction and region merging, with the optimal result selected by an unsupervised index, Q. After the segmentation, a pretrained residual network is fine-tuned on the augmented roof segments using a proposed k-pixel extension technique for classification. The effectiveness of the proposed algorithm was demonstrated on both HR and UHR imagery collected by EagleView over different study sites. The proposed algorithm has yielded promising results and has outperformed traditional machine learning methods using hand-crafted features.
A comparison of LANDSAT TM to MSS imagery for detecting submerged aquatic vegetation in lower Chesapeake Bay

NASA Technical Reports Server (NTRS)

Ackleson, S. G.; Klemas, V.

1985-01-01

LANDSAT Thematic Mapper (TM) and Multispectral Scanner (MSS) imagery generated simultaneously over Guinea Marsh, Virginia, are assessed in the ability to detect submerged aquatic, bottom-adhering plant canopies (SAV). An unsupervised clustering algorithm is applied to both image types and the resulting classifications compared to SAV distributions derived from color aerial photography. Class confidence and accuracy are first computed for all water areas and then only shallow areas where water depth is less than 6 feet. In both the TM and MSS imagery, masking water areas deeper than 6 ft. resulted in greater classification accuracy at confidence levels greater than 50%. Both systems perform poorly in detecting SAV with crown cover densities less than 70%. On the basis of the spectral resolution, radiometric sensitivity, and location of visible bands, TM imagery does not offer a significant advantage over MSS data for detecting SAV in Lower Chesapeake Bay. However, because the TM imagery represents a higher spatial resolution, smaller SAV canopies may be detected than is possible with MSS data.
A Fast Implementation of the ISOCLUS Algorithm

NASA Technical Reports Server (NTRS)

Memarsadeghi, Nargess; Mount, David M.; Netanyahu, Nathan S.; LeMoigne, Jacqueline

2003-01-01

Unsupervised clustering is a fundamental tool in numerous image processing and remote sensing applications. For example, unsupervised clustering is often used to obtain vegetation maps of an area of interest. This approach is useful when reliable training data are either scarce or expensive, and when relatively little a priori information about the data is available. Unsupervised clustering methods play a significant role in the pursuit of unsupervised classification. One of the most popular and widely used clustering schemes for remote sensing applications is the ISOCLUS algorithm, which is based on the ISODATA method. The algorithm is given a set of n data points (or samples) in d-dimensional space, an integer k indicating the initial number of clusters, and a number of additional parameters. The general goal is to compute a set of cluster centers in d-space. Although there is no specific optimization criterion, the algorithm is similar in spirit to the well known k-means clustering method in which the objective is to minimize the average squared distance of each point to its nearest center, called the average distortion. One significant feature of ISOCLUS over k-means is that clusters may be merged or split, and so the final number of clusters may be different from the number k supplied as part of the input. This algorithm will be described in later in this paper. The ISOCLUS algorithm can run very slowly, particularly on large data sets. Given its wide use in remote sensing, its efficient computation is an important goal. We have developed a fast implementation of the ISOCLUS algorithm. Our improvement is based on a recent acceleration to the k-means algorithm, the filtering algorithm, by Kanungo et al.. They showed that, by storing the data in a kd-tree, it was possible to significantly reduce the running time of k-means. We have adapted this method for the ISOCLUS algorithm. For technical reasons, which are explained later, it is necessary to make a minor modification to the ISOCLUS specification. We provide empirical evidence, on both synthetic and Landsat image data sets, that our algorithm's performance is essentially the same as that of ISOCLUS, but with significantly lower running times. We show that our algorithm runs from 3 to 30 times faster than a straightforward implementation of ISOCLUS. Our adaptation of the filtering algorithm involves the efficient computation of a number of cluster statistics that are needed for ISOCLUS, but not for k-means.
Combining fuzzy set theory and nonlinear stretching enhancement for unsupervised classification of cotton root rot

USDA-ARS?s Scientific Manuscript database

Cotton root rot is a destructive disease affecting cotton production. Accurate identification of infected areas within fields is useful for cost-effective control of the disease. The uncertainties caused by various infection stages and newly infected plants make it difficult to achieve accurate clas...
Collected Notes on the Workshop for Pattern Discovery in Large Databases

NASA Technical Reports Server (NTRS)

Buntine, Wray (Editor); Delalto, Martha (Editor)

1991-01-01

These collected notes are a record of material presented at the Workshop. The core data analysis is addressed that have traditionally required statistical or pattern recognition techniques. Some of the core tasks include classification, discrimination, clustering, supervised and unsupervised learning, discovery and diagnosis, i.e., general pattern discovery.
Wetland Mapping with Quad-Pol Data Acquired during Tandem-X Science Phase

NASA Astrophysics Data System (ADS)

Mleczko, M.; Mroz, M.; Fitrzyk, M.

2016-06-01

The aim of this study was to exploit fully polarimetric SAR data acquired during TanDEM-X - Science Phase (2014/2015) over herbaceous wetlands of the Biebrza National Park (BbNP) in North-Eastern Poland for mapping seasonally flooded grasslands and permanent natural vegetation associations. The main goal of this work was to estimate the advantage of fully polarimetric radar images (QuadPol) versus alternative polarization (AltPol) modes. The methodology consisted in processing of several data subsets through polarimetric decompositions of complex quad-pol datasets, classification of multitemporal backscattering images, complementing backscattering images with Shannon Entropy, exploitation of interferometric coherence from tandem operations. In each case the multidimensional stack of images has been classified using ISODATA unsupervised clustering algorithm. With 6 QUAD-POL TSX/TDX acquisitions it was possible to distinguish correctly 5 thematic classes related to their water regime: permanent water bodies, temporarily flooded areas, wet grasslands, dry grasslands and common reed. This last category was possible to distinguish from deciduous forest only with Yamaguchi 4 component decomposition. The interferometric coherence calculated for tandem pairs turned out not so efficient as expected for this wetland mapping.
Exploiting range imagery: techniques and applications

NASA Astrophysics Data System (ADS)

Armbruster, Walter

2009-07-01

Practically no applications exist for which automatic processing of 2D intensity imagery can equal human visual perception. This is not the case for range imagery. The paper gives examples of 3D laser radar applications, for which automatic data processing can exceed human visual cognition capabilities and describes basic processing techniques for attaining these results. The examples are drawn from the fields of helicopter obstacle avoidance, object detection in surveillance applications, object recognition at high range, multi-object-tracking, and object re-identification in range image sequences. Processing times and recognition performances are summarized. The techniques used exploit the bijective continuity of the imaging process as well as its independence of object reflectivity, emissivity and illumination. This allows precise formulations of the probability distributions involved in figure-ground segmentation, feature-based object classification and model based object recognition. The probabilistic approach guarantees optimal solutions for single images and enables Bayesian learning in range image sequences. Finally, due to recent results in 3D-surface completion, no prior model libraries are required for recognizing and re-identifying objects of quite general object categories, opening the way to unsupervised learning and fully autonomous cognitive systems.
Satellite image based methods for fuels maps updating

NASA Astrophysics Data System (ADS)

Alonso-Benito, Alfonso; Hernandez-Leal, Pedro A.; Arbelo, Manuel; Gonzalez-Calvo, Alejandro; Moreno-Ruiz, Jose A.; Garcia-Lazaro, Jose R.

2016-10-01

Regular updating of fuels maps is important for forest fire management. Nevertheless complex and time consuming field work is usually necessary for this purpose, which prevents a more frequent update. That is why the assessment of the usefulness of satellite data and the development of remote sensing techniques that enable the automatic updating of these maps, is of vital interest. In this work, we have tested the use of the spectral bands of OLI (Operational Land Imager) sensor on board Landsat 8 satellite, for updating the fuels map of El Hierro Island (Spain). From previously digitized map, a set of 200 reference plots for different fuel types was created. A 50% of the plots were randomly used as a training set and the rest were considered for validation. Six supervised and 2 unsupervised classification methods were applied, considering two levels of detail. A first level with only 5 classes (Meadow, Brushwood, Undergrowth canopy cover >50%, Undergrowth canopy cover <15%, and Xeric formations), and the second one containing 19 fuel types. The level 1 classification methods yielded an overall accuracy ranging from 44% for Parellelepided to an 84% for Maximun Likelihood. Meanwhile, level 2 results showed at best, an unacceptable overall accuracy of 34%, which prevents the use of this data for such a detailed characterization. Anyway it has been demonstrated that in some conditions, images of medium spatial resolution, like Landsat 8-OLI, could be a valid tool for an automatic upgrade of fuels maps, minimizing costs and complementing traditional methodologies.
Identification and tracking of vertebrae in ultrasound using deep networks with unsupervised feature learning

NASA Astrophysics Data System (ADS)

Hetherington, Jorden; Pesteie, Mehran; Lessoway, Victoria A.; Abolmaesumi, Purang; Rohling, Robert N.

2017-03-01

Percutaneous needle insertion procedures on the spine often require proper identification of the vertebral level in order to effectively deliver anesthetics and analgesic agents to achieve adequate block. For example, in obstetric epidurals, the target is at the L3-L4 intervertebral space. The current clinical method involves "blind" identification of the vertebral level through manual palpation of the spine, which has only 30% accuracy. This implies the need for better anatomical identification prior to needle insertion. A system is proposed to identify the vertebrae, assigning them to their respective levels, and track them in a standard sequence of ultrasound images, when imaged in the paramedian plane. Machine learning techniques are developed to identify discriminative features of the laminae. In particular, a deep network is trained to automatically learn the anatomical features of the lamina peaks, and classify image patches, for pixel-level classification. The chosen network utilizes multiple connected auto-encoders to learn the anatomy. Pre-processing with ultrasound bone enhancement techniques is done to aid the pixel-level classification performance. Once the lamina are identified, vertebrae are assigned levels and tracked in sequential frames. Experimental results were evaluated against an expert sonographer. Based on data acquired from 15 subjects, vertebrae identification with sensitivity of 95% and precision of 95% was achieved within each frame. Between pairs of subsequently analyzed frames, matches of predicted vertebral level labels were correct in 94% of cases, when compared to matches of manually selected labels
Segmentation, Splitting, and Classification of Overlapping Bacteria in Microscope Images for Automatic Bacterial Vaginosis Diagnosis.

PubMed

Song, Youyi; He, Liang; Zhou, Feng; Chen, Siping; Ni, Dong; Lei, Baiying; Wang, Tianfu

2017-07-01

Quantitative analysis of bacterial morphotypes in the microscope images plays a vital role in diagnosis of bacterial vaginosis (BV) based on the Nugent score criterion. However, there are two main challenges for this task: 1) It is quite difficult to identify the bacterial regions due to various appearance, faint boundaries, heterogeneous shapes, low contrast with the background, and small bacteria sizes with regards to the image. 2) There are numerous bacteria overlapping each other, which hinder us to conduct accurate analysis on individual bacterium. To overcome these challenges, we propose an automatic method in this paper to diagnose BV by quantitative analysis of bacterial morphotypes, which consists of a three-step approach, i.e., bacteria regions segmentation, overlapping bacteria splitting, and bacterial morphotypes classification. Specifically, we first segment the bacteria regions via saliency cut, which simultaneously evaluates the global contrast and spatial weighted coherence. And then Markov random field model is applied for high-quality unsupervised segmentation of small object. We then decompose overlapping bacteria clumps into markers, and associate a pixel with markers to identify evidence for eventual individual bacterium splitting. Next, we extract morphotype features from each bacterium to learn the descriptors and to characterize the types of bacteria using an Adaptive Boosting machine learning framework. Finally, BV diagnosis is implemented based on the Nugent score criterion. Experiments demonstrate that our proposed method achieves high accuracy and efficiency in computation for BV diagnosis.
Multi-scale investigation of shrub encroachment in southern Africa

NASA Astrophysics Data System (ADS)

Aplin, Paul; Marston, Christopher; Wilkinson, David; Field, Richard; O'Regan, Hannah

2016-04-01

There is growing speculation that savannah environments throughout Africa have been subject to shrub encroachment in recent years, whereby grassland is lost to woody vegetation cover. Changes in the relative proportions of grassland and woodland are important in the context of conservation of savannah systems, with implications for faunal distributions, environmental management and tourism. Here, we focus on southern Kruger National Park, South Africa, and investigate whether or not shrub encroachment has occurred over the last decade and a half. We use a multi-scale approach, examining the complementarity of medium (e.g. Landsat TM and OLI) and fine (e.g. QuickBird and WorldView-2) spatial resolution satellite sensor imagery, supported by intensive field survey in 2002 and 2014. We employ semi-automated land cover classification, involving a hybrid unsupervised clustering approach with manual class grouping and checking, followed by change detection post-classification comparison analysis. The results show that shrub encroachment is indeed occurring, a finding evidenced through three fine resolution replicate images plus medium resolution imagery. The results also demonstrate the complementarity of medium and fine resolution imagery, though some thematic information must be sacrificed to maintain high medium resolution classification accuracy. Finally, the findings have broader implications for issues such as vegetation seasonality, spatial transferability and management practices.
Unsupervised seismic facies analysis with spatial constraints using regularized fuzzy c-means

NASA Astrophysics Data System (ADS)

Song, Chengyun; Liu, Zhining; Cai, Hanpeng; Wang, Yaojun; Li, Xingming; Hu, Guangmin

2017-12-01

Seismic facies analysis techniques combine classification algorithms and seismic attributes to generate a map that describes main reservoir heterogeneities. However, most of the current classification algorithms only view the seismic attributes as isolated data regardless of their spatial locations, and the resulting map is generally sensitive to noise. In this paper, a regularized fuzzy c-means (RegFCM) algorithm is used for unsupervised seismic facies analysis. Due to the regularized term of the RegFCM algorithm, the data whose adjacent locations belong to same classification will play a more important role in the iterative process than other data. Therefore, this method can reduce the effect of seismic data noise presented in discontinuous regions. The synthetic data with different signal/noise values are used to demonstrate the noise tolerance ability of the RegFCM algorithm. Meanwhile, the fuzzy factor, the neighbour window size and the regularized weight are tested using various values, to provide a reference of how to set these parameters. The new approach is also applied to a real seismic data set from the F3 block of the Netherlands. The results show improved spatial continuity, with clear facies boundaries and channel morphology, which reveals that the method is an effective seismic facies analysis tool.

Unsupervised segmentation of low-contrast multichannel images: discrimination of tissue components in microscopic images of unstained specimens

NASA Astrophysics Data System (ADS)

Kopriva, Ivica; Popović Hadžija, Marijana; Hadžija, Mirko; Aralica, Gorana

2015-06-01

Low-contrast images, such as color microscopic images of unstained histological specimens, are composed of objects with highly correlated spectral profiles. Such images are very hard to segment. Here, we present a method that nonlinearly maps low-contrast color image into an image with an increased number of non-physical channels and a decreased correlation between spectral profiles. The method is a proof-of-concept validated on the unsupervised segmentation of color images of unstained specimens, in which case the tissue components appear colorless when viewed under the light microscope. Specimens of human hepatocellular carcinoma, human liver with metastasis from colon and gastric cancer and mouse fatty liver were used for validation. The average correlation between the spectral profiles of the tissue components was greater than 0.9985, and the worst case correlation was greater than 0.9997. The proposed method can potentially be applied to the segmentation of low-contrast multichannel images with high spatial resolution that arise in other imaging modalities.
Evaluating suitability of Pol-SAR (TerraSAR-X, Radarsat-2) for automated sea ice classification

NASA Astrophysics Data System (ADS)

Ressel, Rudolf; Singha, Suman; Lehner, Susanne

2016-05-01

Satellite borne SAR imagery has become an invaluable tool in the field of sea ice monitoring. Previously, single polarimetric imagery were employed in supervised and unsupervised classification schemes for sea ice investigation, which was preceded by image processing techniques such as segmentation and textural features. Recently, through the advent of polarimetric SAR sensors, investigation of polarimetric features in sea ice has attracted increased attention. While dual-polarimetric data has already been investigated in a number of works, full-polarimetric data has so far not been a major scientific focus. To explore the possibilities of full-polarimetric data and compare the differences in C- and X-bands, we endeavor to analyze in detail an array of datasets, simultaneously acquired, in C-band (RADARSAT-2) and X-band (TerraSAR-X) over ice infested areas. First, we propose an array of polarimetric features (Pauli and lexicographic based). Ancillary data from national ice services, SMOS data and expert judgement were utilized to identify the governing ice regimes. Based on these observations, we then extracted mentioned features. The subsequent supervised classification approach was based on an Artificial Neural Network (ANN). To gain quantitative insight into the quality of the features themselves (and reduce a possible impact of the Hughes phenomenon), we employed mutual information to unearth the relevance and redundancy of features. The results of this information theoretic analysis guided a pruning process regarding the optimal subset of features. In the last step we compared the classified results of all sensors and images, stated respective accuracies and discussed output discrepancies in the cases of simultaneous acquisitions.
Mapping of rock types using a joint approach by combining the multivariate statistics, self-organizing map and Bayesian neural networks: an example from IODP 323 site

NASA Astrophysics Data System (ADS)

Karmakar, Mampi; Maiti, Saumen; Singh, Amrita; Ojha, Maheswar; Maity, Bhabani Sankar

2017-07-01

Modeling and classification of the subsurface lithology is very important to understand the evolution of the earth system. However, precise classification and mapping of lithology using a single framework are difficult due to the complexity and the nonlinearity of the problem driven by limited core sample information. Here, we implement a joint approach by combining the unsupervised and the supervised methods in a single framework for better classification and mapping of rock types. In the unsupervised method, we use the principal component analysis (PCA), K-means cluster analysis (K-means), dendrogram analysis, Fuzzy C-means (FCM) cluster analysis and self-organizing map (SOM). In the supervised method, we use the Bayesian neural networks (BNN) optimized by the Hybrid Monte Carlo (HMC) (BNN-HMC) and the scaled conjugate gradient (SCG) (BNN-SCG) techniques. We use P-wave velocity, density, neutron porosity, resistivity and gamma ray logs of the well U1343E of the Integrated Ocean Drilling Program (IODP) Expedition 323 in the Bering Sea slope region. While the SOM algorithm allows us to visualize the clustering results in spatial domain, the combined classification schemes (supervised and unsupervised) uncover the different patterns of lithology such of as clayey-silt, diatom-silt and silty-clay from an un-cored section of the drilled hole. In addition, the BNN approach is capable of estimating uncertainty in the predictive modeling of three types of rocks over the entire lithology section at site U1343. Alternate succession of clayey-silt, diatom-silt and silty-clay may be representative of crustal inhomogeneity in general and thus could be a basis for detail study related to the productivity of methane gas in the oceans worldwide. Moreover, at the 530 m depth down below seafloor (DSF), the transition from Pliocene to Pleistocene could be linked to lithological alternation between the clayey-silt and the diatom-silt. The present results could provide the basis for the detailed study to get deeper insight into the Bering Sea' sediment deposition and sequence.
Performance analysis of unsupervised optimal fuzzy clustering algorithm for MRI brain tumor segmentation.

PubMed

Blessy, S A Praylin Selva; Sulochana, C Helen

2015-01-01

Segmentation of brain tumor from Magnetic Resonance Imaging (MRI) becomes very complicated due to the structural complexities of human brain and the presence of intensity inhomogeneities. To propose a method that effectively segments brain tumor from MR images and to evaluate the performance of unsupervised optimal fuzzy clustering (UOFC) algorithm for segmentation of brain tumor from MR images. Segmentation is done by preprocessing the MR image to standardize intensity inhomogeneities followed by feature extraction, feature fusion and clustering. Different validation measures are used to evaluate the performance of the proposed method using different clustering algorithms. The proposed method using UOFC algorithm produces high sensitivity (96%) and low specificity (4%) compared to other clustering methods. Validation results clearly show that the proposed method with UOFC algorithm effectively segments brain tumor from MR images.
A Comparative Evaluation of Unsupervised Anomaly Detection Algorithms for Multivariate Data

PubMed Central

Goldstein, Markus; Uchida, Seiichi

2016-01-01

Anomaly detection is the process of identifying unexpected items or events in datasets, which differ from the norm. In contrast to standard classification tasks, anomaly detection is often applied on unlabeled data, taking only the internal structure of the dataset into account. This challenge is known as unsupervised anomaly detection and is addressed in many practical applications, for example in network intrusion detection, fraud detection as well as in the life science and medical domain. Dozens of algorithms have been proposed in this area, but unfortunately the research community still lacks a comparative universal evaluation as well as common publicly available datasets. These shortcomings are addressed in this study, where 19 different unsupervised anomaly detection algorithms are evaluated on 10 different datasets from multiple application domains. By publishing the source code and the datasets, this paper aims to be a new well-funded basis for unsupervised anomaly detection research. Additionally, this evaluation reveals the strengths and weaknesses of the different approaches for the first time. Besides the anomaly detection performance, computational effort, the impact of parameter settings as well as the global/local anomaly detection behavior is outlined. As a conclusion, we give an advise on algorithm selection for typical real-world tasks. PMID:27093601
Detection of food intake from swallowing sequences by supervised and unsupervised methods.

PubMed

Lopez-Meyer, Paulo; Makeyev, Oleksandr; Schuckers, Stephanie; Melanson, Edward L; Neuman, Michael R; Sazonov, Edward

2010-08-01

Studies of food intake and ingestive behavior in free-living conditions most often rely on self-reporting-based methods that can be highly inaccurate. Methods of Monitoring of Ingestive Behavior (MIB) rely on objective measures derived from chewing and swallowing sequences and thus can be used for unbiased study of food intake with free-living conditions. Our previous study demonstrated accurate detection of food intake in simple models relying on observation of both chewing and swallowing. This article investigates methods that achieve comparable accuracy of food intake detection using only the time series of swallows and thus eliminating the need for the chewing sensor. The classification is performed for each individual swallow rather than for previously used time slices and thus will lead to higher accuracy in mass prediction models relying on counts of swallows. Performance of a group model based on a supervised method (SVM) is compared to performance of individual models based on an unsupervised method (K-means) with results indicating better performance of the unsupervised, self-adapting method. Overall, the results demonstrate that highly accurate detection of intake of foods with substantially different physical properties is possible by an unsupervised system that relies on the information provided by the swallowing alone.
Detection of Food Intake from Swallowing Sequences by Supervised and Unsupervised Methods

PubMed Central

Lopez-Meyer, Paulo; Makeyev, Oleksandr; Schuckers, Stephanie; Melanson, Edward L.; Neuman, Michael R.; Sazonov, Edward

2010-01-01

Studies of food intake and ingestive behavior in free-living conditions most often rely on self-reporting-based methods that can be highly inaccurate. Methods of Monitoring of Ingestive Behavior (MIB) rely on objective measures derived from chewing and swallowing sequences and thus can be used for unbiased study of food intake with free-living conditions. Our previous study demonstrated accurate detection of food intake in simple models relying on observation of both chewing and swallowing. This article investigates methods that achieve comparable accuracy of food intake detection using only the time series of swallows and thus eliminating the need for the chewing sensor. The classification is performed for each individual swallow rather than for previously used time slices and thus will lead to higher accuracy in mass prediction models relying on counts of swallows. Performance of a group model based on a supervised method (SVM) is compared to performance of individual models based on an unsupervised method (K-means) with results indicating better performance of the unsupervised, self-adapting method. Overall, the results demonstrate that highly accurate detection of intake of foods with substantially different physical properties is possible by an unsupervised system that relies on the information provided by the swallowing alone. PMID:20352335
Metric Learning to Enhance Hyperspectral Image Segmentation

NASA Technical Reports Server (NTRS)

Thompson, David R.; Castano, Rebecca; Bue, Brian; Gilmore, Martha S.

2013-01-01

Unsupervised hyperspectral image segmentation can reveal spatial trends that show the physical structure of the scene to an analyst. They highlight borders and reveal areas of homogeneity and change. Segmentations are independently helpful for object recognition, and assist with automated production of symbolic maps. Additionally, a good segmentation can dramatically reduce the number of effective spectra in an image, enabling analyses that would otherwise be computationally prohibitive. Specifically, using an over-segmentation of the image instead of individual pixels can reduce noise and potentially improve the results of statistical post-analysis. In this innovation, a metric learning approach is presented to improve the performance of unsupervised hyperspectral image segmentation. The prototype demonstrations attempt a superpixel segmentation in which the image is conservatively over-segmented; that is, the single surface features may be split into multiple segments, but each individual segment, or superpixel, is ensured to have homogenous mineralogy.
Interpretation of Pennsylvania agricultural land use from ERTS-1 data

NASA Technical Reports Server (NTRS)

Mcmurtry, G. J.; Petersen, G. W. (Principal Investigator); Wilson, A. D.

1974-01-01

The author has identified the following significant results. To study the complex agricultural patterns in Pennsylvania, a portion of an ERTS scene was selected for detailed analysis. Various photographic products were made and were found to be only of limited value. This necessitated the digital processing of the ERTS data. Using an unsupervised classification procedure, it was possible to delineate the following categories: (1) forest land with a northern aspect, (2) forest land with a southern aspect, (3) valley trees, (4) wheat, (5) corn, (6) alfalfa, grass, pasture, (7) disturbed land, (8) builtup land, (9) strip mines, and (10) water. These land use categories were delineated at a scale of approximately 1:20,000 on the line printer output. Land use delineations were also made using the General Electric IMAGE 100 interactive analysis system.
Polarimetry-Based Land Cover Classification with Sentinel-1 Data

NASA Astrophysics Data System (ADS)

Banque, Xavier; Lopez-Sanchez, Juan M.; Monells, Daniel; Ballester, David; Duro, Javier; Koudogbo, Fifame

2015-04-01

The presented research focuses on the assessment of the exploitation of the Sentinel-1 dual polarization data for land cover classification. In order to take advantage of massive data availability produced by Sentinel-1, data used in this research work is Interferometric Wide Swath mode, acquired over the Altmühlsee, Weißenburg-Gunzenhausen, Germany during November 2014. The developed preliminary classifier is based on the interpretation of several polarimetric figures as well as the Dual Polarization Entropy/Alpha Decomposition. Specifically, the following polarimetric indicators will be assessed: the channels cross- correlation, the cross and co-polar channels ratio and both cross and co-polar backscattering coefficients. The work carried out concentrates on the joint interpretation of the backscattering response of the co-pol and cross- pol channels for four or five different distributed targets that set the basis for an unsupervised simple land cover classifier. The developed research targets a preliminary unsupervised classifier able to differentiate between four or five terrain classes, including water, urban, forest and bare soil. Obtained results pave the way for the development of a Sentinel-1 based land classifier.
Knee cartilage extraction and bone-cartilage interface analysis from 3D MRI data sets

NASA Astrophysics Data System (ADS)

Tamez-Pena, Jose G.; Barbu-McInnis, Monica; Totterman, Saara

2004-05-01

This works presents a robust methodology for the analysis of the knee joint cartilage and the knee bone-cartilage interface from fused MRI sets. The proposed approach starts by fusing a set of two 3D MR images the knee. Although the proposed method is not pulse sequence dependent, the first sequence should be programmed to achieve good contrast between bone and cartilage. The recommended second pulse sequence is one that maximizes the contrast between cartilage and surrounding soft tissues. Once both pulse sequences are fused, the proposed bone-cartilage analysis is done in four major steps. First, an unsupervised segmentation algorithm is used to extract the femur, the tibia, and the patella. Second, a knowledge based feature extraction algorithm is used to extract the femoral, tibia and patellar cartilages. Third, a trained user corrects cartilage miss-classifications done by the automated extracted cartilage. Finally, the final segmentation is the revisited using an unsupervised MAP voxel relaxation algorithm. This final segmentation has the property that includes the extracted bone tissue as well as all the cartilage tissue. This is an improvement over previous approaches where only the cartilage was segmented. Furthermore, this approach yields very reproducible segmentation results in a set of scan-rescan experiments. When these segmentations were coupled with a partial volume compensated surface extraction algorithm the volume, area, thickness measurements shows precisions around 2.6%
Unsupervised Feature Selection Based on the Morisita Index for Hyperspectral Images

NASA Astrophysics Data System (ADS)

Golay, Jean; Kanevski, Mikhail

2017-04-01

Hyperspectral sensors are capable of acquiring images with hundreds of narrow and contiguous spectral bands. Compared with traditional multispectral imagery, the use of hyperspectral images allows better performance in discriminating between land-cover classes, but it also results in large redundancy and high computational data processing. To alleviate such issues, unsupervised feature selection techniques for redundancy minimization can be implemented. Their goal is to select the smallest subset of features (or bands) in such a way that all the information content of a data set is preserved as much as possible. The present research deals with the application to hyperspectral images of a recently introduced technique of unsupervised feature selection: the Morisita-Based filter for Redundancy Minimization (MBRM). MBRM is based on the (multipoint) Morisita index of clustering and on the Morisita estimator of Intrinsic Dimension (ID). The fundamental idea of the technique is to retain only the bands which contribute to increasing the ID of an image. In this way, redundant bands are disregarded, since they have no impact on the ID. Besides, MBRM has several advantages over benchmark techniques: in addition to its ability to deal with large data sets, it can capture highly-nonlinear dependences and its implementation is straightforward in any programming environment. Experimental results on freely available hyperspectral images show the good effectiveness of MBRM in remote sensing data processing. Comparisons with benchmark techniques are carried out and random forests are used to assess the performance of MBRM in reducing the data dimensionality without loss of relevant information. References [1] C. Traina Jr., A.J.M. Traina, L. Wu, C. Faloutsos, Fast feature selection using fractal dimension, in: Proceedings of the XV Brazilian Symposium on Databases, SBBD, pp. 158-171, 2000. [2] J. Golay, M. Kanevski, A new estimator of intrinsic dimension based on the multipoint Morisita index, Pattern Recognition 48(12), pp. 4070-4081, 2015. [3] J. Golay, M. Kanevski, Unsupervised feature selection based on the Morisita estimator of intrinsic dimension, arXiv:1608.05581, 2016.
Back-island and open-ocean shorelines, and sand areas of Assateague Island, Maryland and Virginia, April 12, 1989, to September 5, 2013

USGS Publications Warehouse

Guy, Kristy K.

2015-01-01

This Data Series Report includes several open-ocean shorelines, back-island shorelines, back-island shoreline points, sand area polygons, and sand lines for Assateague Island that were extracted from natural-color orthoimagery (aerial photography) dated from April 12, 1989, to September 5, 2013. The images used were 0.3–2-meter (m)-resolution U.S. Geological Survey Digital Orthophoto Quarter Quads (DOQQ), U.S. Department of Agriculture National Agriculture Imagery Program (NAIP) images, and Virginia Geographic Information Network Virginia Base Map Program (VBMP) images courtesy of the Commonwealth of Virginia. The back-island shorelines were hand-digitized at the intersect of the apparent back-island shoreline and transects spaced at 20-m intervals. The open-ocean shorelines were hand-digitized at the approximate still water level, such as tide level, which was fit through the average position of waves and swash apparent on the beach. Hand-digitizing was done at a scale of approximately 1:2,000. The sand polygons were derived by using an image-processing unsupervised classification technique that separates images into classes. The classes were then visually categorized as either sand or not sand. Also included in this report are 20-m-spaced transect lines and the transect base lines.
An Introduction to Topic Modeling as an Unsupervised Machine Learning Way to Organize Text Information

ERIC Educational Resources Information Center

Snyder, Robin M.

2015-01-01

The field of topic modeling has become increasingly important over the past few years. Topic modeling is an unsupervised machine learning way to organize text (or image or DNA, etc.) information such that related pieces of text can be identified. This paper/session will present/discuss the current state of topic modeling, why it is important, and…
Wire connector classification with machine vision and a novel hybrid SVM

NASA Astrophysics Data System (ADS)

Chauhan, Vedang; Joshi, Keyur D.; Surgenor, Brian W.

2018-04-01

A machine vision-based system has been developed and tested that uses a novel hybrid Support Vector Machine (SVM) in a part inspection application with clear plastic wire connectors. The application required the system to differentiate between 4 different known styles of connectors plus one unknown style, for a total of 5 classes. The requirement to handle an unknown class is what necessitated the hybrid approach. The system was trained with the 4 known classes and tested with 5 classes (the 4 known plus the 1 unknown). The hybrid classification approach used two layers of SVMs: one layer was semi-supervised and the other layer was supervised. The semi-supervised SVM was a special case of unsupervised machine learning that classified test images as one of the 4 known classes (to accept) or as the unknown class (to reject). The supervised SVM classified test images as one of the 4 known classes and consequently would give false positives (FPs). Two methods were tested. The difference between the methods was that the order of the layers was switched. The method with the semi-supervised layer first gave an accuracy of 80% with 20% FPs. The method with the supervised layer first gave an accuracy of 98% with 0% FPs. Further work is being conducted to see if the hybrid approach works with other applications that have an unknown class requirement.
Biologically Inspired Model for Visual Cognition Achieving Unsupervised Episodic and Semantic Feature Learning.

PubMed

Qiao, Hong; Li, Yinlin; Li, Fengfu; Xi, Xuanyang; Wu, Wei

2016-10-01

Recently, many biologically inspired visual computational models have been proposed. The design of these models follows the related biological mechanisms and structures, and these models provide new solutions for visual recognition tasks. In this paper, based on the recent biological evidence, we propose a framework to mimic the active and dynamic learning and recognition process of the primate visual cortex. From principle point of view, the main contributions are that the framework can achieve unsupervised learning of episodic features (including key components and their spatial relations) and semantic features (semantic descriptions of the key components), which support higher level cognition of an object. From performance point of view, the advantages of the framework are as follows: 1) learning episodic features without supervision-for a class of objects without a prior knowledge, the key components, their spatial relations and cover regions can be learned automatically through a deep neural network (DNN); 2) learning semantic features based on episodic features-within the cover regions of the key components, the semantic geometrical values of these components can be computed based on contour detection; 3) forming the general knowledge of a class of objects-the general knowledge of a class of objects can be formed, mainly including the key components, their spatial relations and average semantic values, which is a concise description of the class; and 4) achieving higher level cognition and dynamic updating-for a test image, the model can achieve classification and subclass semantic descriptions. And the test samples with high confidence are selected to dynamically update the whole model. Experiments are conducted on face images, and a good performance is achieved in each layer of the DNN and the semantic description learning process. Furthermore, the model can be generalized to recognition tasks of other objects with learning ability.
Semi-automated surface mapping via unsupervised classification

NASA Astrophysics Data System (ADS)

D'Amore, M.; Le Scaon, R.; Helbert, J.; Maturilli, A.

2017-09-01

Due to the increasing volume of the returned data from space mission, the human search for correlation and identification of interesting features becomes more and more unfeasible. Statistical extraction of features via machine learning methods will increase the scientific output of remote sensing missions and aid the discovery of yet unknown feature hidden in dataset. Those methods exploit algorithm trained on features from multiple instrument, returning classification maps that explore intra-dataset correlation, allowing for the discovery of unknown features. We present two applications, one for Mercury and one for Vesta.
Statewide land cover derived from multiseasonal Landsat TM data: A retrospective of the WISCLAND project

USGS Publications Warehouse

Reese, H.M.; Lillesand, T.M.; Nagel, D.E.; Stewart, J.S.; Goldmann, R.A.; Simmons, T.E.; Chipman, J.W.; Tessar, P.A.

2002-01-01

Landsat Thematic Mapper (TM) data were the basis in production of a statewide land cover data set for Wisconsin, undertaken in partnership with U.S. Geological Survey's (USGS) Gap Analysis Program (GAP). The data set contained seven classes comparable to Anderson Level I and 24 classes comparable to Anderson Level II/III. Twelve scenes of dual-date TM data were processed with methods that included principal components analysis, stratification into spectrally consistent units, separate classification of upland, wetland, and urban areas, and a hybrid supervised/unsupervised classification called "guided clustering." The final data had overall accuracies of 94% for Anderson Level I upland classes, 77% for Level II/III upland classes, and 84% for Level II/III wetland classes. Classification accuracies for deciduous and coniferous forest were 95% and 93%, respectively, and forest species' overall accuracies ranged from 70% to 84%. Limited availability of acceptable imagery necessitated use of an early May date in a majority of scene pairs, perhaps contributing to lower accuracy for upland deciduous forest species. The mixed deciduous/coniferous forest class had the lowest accuracy, most likely due to distinctly classifying a purely mixed class. Mixed forest signatures containing oak were often confused with pure oak. Guided clustering was seen as an efficient classification method, especially at the tree species level, although its success relied in part on image dates, accurate ground troth, and some analyst intervention. ?? 2002 Elsevier Science Inc. All rights reserved.
Classification Accuracy Increase Using Multisensor Data Fusion

NASA Astrophysics Data System (ADS)

Makarau, A.; Palubinskas, G.; Reinartz, P.

2011-09-01

The practical use of very high resolution visible and near-infrared (VNIR) data is still growing (IKONOS, Quickbird, GeoEye-1, etc.) but for classification purposes the number of bands is limited in comparison to full spectral imaging. These limitations may lead to the confusion of materials such as different roofs, pavements, roads, etc. and therefore may provide wrong interpretation and use of classification products. Employment of hyperspectral data is another solution, but their low spatial resolution (comparing to multispectral data) restrict their usage for many applications. Another improvement can be achieved by fusion approaches of multisensory data since this may increase the quality of scene classification. Integration of Synthetic Aperture Radar (SAR) and optical data is widely performed for automatic classification, interpretation, and change detection. In this paper we present an approach for very high resolution SAR and multispectral data fusion for automatic classification in urban areas. Single polarization TerraSAR-X (SpotLight mode) and multispectral data are integrated using the INFOFUSE framework, consisting of feature extraction (information fission), unsupervised clustering (data representation on a finite domain and dimensionality reduction), and data aggregation (Bayesian or neural network). This framework allows a relevant way of multisource data combination following consensus theory. The classification is not influenced by the limitations of dimensionality, and the calculation complexity primarily depends on the step of dimensionality reduction. Fusion of single polarization TerraSAR-X, WorldView-2 (VNIR or full set), and Digital Surface Model (DSM) data allow for different types of urban objects to be classified into predefined classes of interest with increased accuracy. The comparison to classification results of WorldView-2 multispectral data (8 spectral bands) is provided and the numerical evaluation of the method in comparison to other established methods illustrates the advantage in the classification accuracy for many classes such as buildings, low vegetation, sport objects, forest, roads, rail roads, etc.
Unsupervised Neural Network Quantifies the Cost of Visual Information Processing.

PubMed

Orbán, Levente L; Chartier, Sylvain

2015-01-01

Untrained, "flower-naïve" bumblebees display behavioural preferences when presented with visual properties such as colour, symmetry, spatial frequency and others. Two unsupervised neural networks were implemented to understand the extent to which these models capture elements of bumblebees' unlearned visual preferences towards flower-like visual properties. The computational models, which are variants of Independent Component Analysis and Feature-Extracting Bidirectional Associative Memory, use images of test-patterns that are identical to ones used in behavioural studies. Each model works by decomposing images of floral patterns into meaningful underlying factors. We reconstruct the original floral image using the components and compare the quality of the reconstructed image to the original image. Independent Component Analysis matches behavioural results substantially better across several visual properties. These results are interpreted to support a hypothesis that the temporal and energetic costs of information processing by pollinators served as a selective pressure on floral displays: flowers adapted to pollinators' cognitive constraints.

Unsupervised Unmixing of Hyperspectral Images Accounting for Endmember Variability.

PubMed

Halimi, Abderrahim; Dobigeon, Nicolas; Tourneret, Jean-Yves

2015-12-01

This paper presents an unsupervised Bayesian algorithm for hyperspectral image unmixing, accounting for endmember variability. The pixels are modeled by a linear combination of endmembers weighted by their corresponding abundances. However, the endmembers are assumed random to consider their variability in the image. An additive noise is also considered in the proposed model, generalizing the normal compositional model. The proposed algorithm exploits the whole image to benefit from both spectral and spatial information. It estimates both the mean and the covariance matrix of each endmember in the image. This allows the behavior of each material to be analyzed and its variability to be quantified in the scene. A spatial segmentation is also obtained based on the estimated abundances. In order to estimate the parameters associated with the proposed Bayesian model, we propose to use a Hamiltonian Monte Carlo algorithm. The performance of the resulting unmixing strategy is evaluated through simulations conducted on both synthetic and real data.
Unsupervised analysis of small animal dynamic Cerenkov luminescence imaging

NASA Astrophysics Data System (ADS)

Spinelli, Antonello E.; Boschi, Federico

2011-12-01

Clustering analysis (CA) and principal component analysis (PCA) were applied to dynamic Cerenkov luminescence images (dCLI). In order to investigate the performances of the proposed approaches, two distinct dynamic data sets obtained by injecting mice with 32P-ATP and 18F-FDG were acquired using the IVIS 200 optical imager. The k-means clustering algorithm has been applied to dCLI and was implemented using interactive data language 8.1. We show that cluster analysis allows us to obtain good agreement between the clustered and the corresponding emission regions like the bladder, the liver, and the tumor. We also show a good correspondence between the time activity curves of the different regions obtained by using CA and manual region of interest analysis on dCLIT and PCA images. We conclude that CA provides an automatic unsupervised method for the analysis of preclinical dynamic Cerenkov luminescence image data.
An Unsupervised Approach for Extraction of Blood Vessels from Fundus Images.

PubMed

Dash, Jyotiprava; Bhoi, Nilamani

2018-04-26

Pathological disorders may happen due to small changes in retinal blood vessels which may later turn into blindness. Hence, the accurate segmentation of blood vessels is becoming a challenging task for pathological analysis. This paper offers an unsupervised recursive method for extraction of blood vessels from ophthalmoscope images. First, a vessel-enhanced image is generated with the help of gamma correction and contrast-limited adaptive histogram equalization (CLAHE). Next, the vessels are extracted iteratively by applying an adaptive thresholding technique. At last, a final vessel segmented image is produced by applying a morphological cleaning operation. Evaluations are accompanied on the publicly available digital retinal images for vessel extraction (DRIVE) and Child Heart And Health Study in England (CHASE_DB1) databases using nine different measurements. The proposed method achieves average accuracies of 0.957 and 0.952 on DRIVE and CHASE_DB1 databases respectively.
Mapping South San Francisco Bay's seabed diversity for use in wetland restoration planning

USGS Publications Warehouse

Fregoso, Theresa A.; Jaffe, B.; Rathwell, G.; Collins, W.; Rhynas, K.; Tomlin, V.; Sullivan, S.

2006-01-01

Data for an acoustic seabed classification were collected as a part of a California Coastal Conservancy funded bathymetric survey of South Bay in early 2005. A QTC VIEW seabed classification system recorded echoes from a sungle bean 50 kHz echosounder. Approximately 450,000 seabed classification records were generated from an are of of about 30 sq. miles. Ten district acoustic classes were identified through an unsupervised classification system using principle component and cluster analyses. One hundred and sixty-one grab samples and forty-five benthic community composition data samples collected in the study area shortly before and after the seabed classification survey, further refined the ten classes into groups based on grain size. A preliminary map of surficial grain size of South Bay was developed from the combination of the seabed classification and the grab and benthic samples. The initial seabed classification map, the grain size map, and locations of sediment samples will be displayed along with the methods of acousitc seabed classification.
Deep neural ensemble for retinal vessel segmentation in fundus images towards achieving label-free angiography.

PubMed

Lahiri, A; Roy, Abhijit Guha; Sheet, Debdoot; Biswas, Prabir Kumar

2016-08-01

Automated segmentation of retinal blood vessels in label-free fundus images entails a pivotal role in computed aided diagnosis of ophthalmic pathologies, viz., diabetic retinopathy, hypertensive disorders and cardiovascular diseases. The challenge remains active in medical image analysis research due to varied distribution of blood vessels, which manifest variations in their dimensions of physical appearance against a noisy background. In this paper we formulate the segmentation challenge as a classification task. Specifically, we employ unsupervised hierarchical feature learning using ensemble of two level of sparsely trained denoised stacked autoencoder. First level training with bootstrap samples ensures decoupling and second level ensemble formed by different network architectures ensures architectural revision. We show that ensemble training of auto-encoders fosters diversity in learning dictionary of visual kernels for vessel segmentation. SoftMax classifier is used for fine tuning each member autoencoder and multiple strategies are explored for 2-level fusion of ensemble members. On DRIVE dataset, we achieve maximum average accuracy of 95.33% with an impressively low standard deviation of 0.003 and Kappa agreement coefficient of 0.708. Comparison with other major algorithms substantiates the high efficacy of our model.
Landsat and Sentinel-2A Surface Albedo Estimation and Evaluation Against In Situ Measurements Across the US SURFRAD Network

NASA Astrophysics Data System (ADS)

Franch, B.; Skakun, S.; Vermote, E.; Roger, J. C.

2017-12-01

Surface albedo is an essential parameter not only for developing climate models, but also for most energy balance studies. While climate models are usually applied at coarse resolution, the energy balance studies, which are mainly focused on agricultural applications, require a high spatial resolution. The albedo, estimated through the angular integration of the BRDF, requires an appropriate angular sampling of the surface. However, Sentinel-2A sampling characteristics, with nearly constant observation geometry and low illumination variation, prevent from deriving a surface albedo product. In this work, we apply an algorithm developed to derive a Landsat surface albedo to Sentinel-2A. It is based on the BRDF parameters estimated from the MODerate Resolution Imaging Spectroradiometer (MODIS) CMG surface reflectance product (M{O,Y}D09) using the VJB method (Vermote et al., 2009). Sentinel-2A unsupervised classification images are used to disaggregate the BRDF parameters to the Sentinel-2 spatial resolution. We test the results over five different sites of the US SURFRAD network and plot the results versus albedo field measurements. Additionally, we also test this methodology using Landsat-8 images.
Improved regional-scale Brazilian cropping systems' mapping based on a semi-automatic object-based clustering approach

NASA Astrophysics Data System (ADS)

Bellón, Beatriz; Bégué, Agnès; Lo Seen, Danny; Lebourgeois, Valentine; Evangelista, Balbino Antônio; Simões, Margareth; Demonte Ferraz, Rodrigo Peçanha

2018-06-01

Cropping systems' maps at fine scale over large areas provide key information for further agricultural production and environmental impact assessments, and thus represent a valuable tool for effective land-use planning. There is, therefore, a growing interest in mapping cropping systems in an operational manner over large areas, and remote sensing approaches based on vegetation index time series analysis have proven to be an efficient tool. However, supervised pixel-based approaches are commonly adopted, requiring resource consuming field campaigns to gather training data. In this paper, we present a new object-based unsupervised classification approach tested on an annual MODIS 16-day composite Normalized Difference Vegetation Index time series and a Landsat 8 mosaic of the State of Tocantins, Brazil, for the 2014-2015 growing season. Two variants of the approach are compared: an hyperclustering approach, and a landscape-clustering approach involving a previous stratification of the study area into landscape units on which the clustering is then performed. The main cropping systems of Tocantins, characterized by the crop types and cropping patterns, were efficiently mapped with the landscape-clustering approach. Results show that stratification prior to clustering significantly improves the classification accuracies for underrepresented and sparsely distributed cropping systems. This study illustrates the potential of unsupervised classification for large area cropping systems' mapping and contributes to the development of generic tools for supporting large-scale agricultural monitoring across regions.
Back-Island and Open-Ocean Shorelines, and Sand Areas of the Undeveloped Areas of New Jersey Barrier Islands, March 9, 1991, to July 30, 2013

USGS Publications Warehouse

Guy, Kristy K.

2015-11-09

This Data Series Report includes open-ocean shorelines, back-island shorelines, back-island shoreline points, sand polygons, and sand lines for the undeveloped areas of New Jersey barrier islands. These data were extracted from orthoimagery (aerial photography) taken between March 9, 1991, and July 30, 2013. The images used were 0.3–1-meter (m)-resolution U.S. Geological Survey Digital Orthophoto Quarter Quads (DOQQ), U.S. Department of Agriculture National Agriculture Imagery Program (NAIP) images, National Oceanic and Atmospheric Administration images, and New Jersey Geographic Information Network images. The back-island shorelines were hand-digitized at the intersects of the apparent back-island shoreline and transects spaced at 20-m intervals. The open-ocean shorelines were hand-digitized at the approximate still-water level, such as tide level, which was fit through the average position of waves and swash apparent on the beach. Hand-digitizing was done at a scale of approximately 1:2,000. The sand polygons were derived by an image-processing unsupervised classification technique that separates images into classes. The classes were then visually categorized as either sand or not sand. Sand lines were taken from the sand polygons. Also included in this report are 20-m-spaced transect lines and the transect base lines.
Anaysis of the quality of image data required by the LANDSAT-4 Thematic Mapper and Multispectral Scanner. [agricultural and forest cover types in California

NASA Technical Reports Server (NTRS)

Colwell, R. N. (Principal Investigator)

1984-01-01

The spatial, geometric, and radiometric qualities of LANDSAT 4 thematic mapper (TM) and multispectral scanner (MSS) data were evaluated by interpreting, through visual and computer means, film and digital products for selected agricultural and forest cover types in California. Multispectral analyses employing Bayesian maximum likelihood, discrete relaxation, and unsupervised clustering algorithms were used to compare the usefulness of TM and MSS data for discriminating individual cover types. Some of the significant results are as follows: (1) for maximizing the interpretability of agricultural and forest resources, TM color composites should contain spectral bands in the visible, near-reflectance infrared, and middle-reflectance infrared regions, namely TM 4 and TM % and must contain TM 4 in all cases even at the expense of excluding TM 5; (2) using enlarged TM film products, planimetric accuracy of mapped poins was within 91 meters (RMSE east) and 117 meters (RMSE north); (3) using TM digital products, planimetric accuracy of mapped points was within 12.0 meters (RMSE east) and 13.7 meters (RMSE north); and (4) applying a contextual classification algorithm to TM data provided classification accuracies competitive with Bayesian maximum likelihood.
Seafloor habitat mapping of the New York Bight incorporating sidescan sonar data

USGS Publications Warehouse

Lathrop, R.G.; Cole, M.; Senyk, N.; Butman, B.

2006-01-01

The efficacy of using sidescan sonar imagery, image classification algorithms and geographic information system (GIS) techniques to characterize the seafloor bottom of the New York Bight were assessed. The resulting seafloor bottom type map was compared with fish trawl survey data to determine whether there were any discernable habitat associations. An unsupervised classification with 20 spectral classes was produced using the sidescan sonar imagery, bathymetry and secondarily derived spatial heterogeneity to characterize homogenous regions within the study area. The spectral classes, geologic interpretations of the study region, bathymetry and a bottom landform index were used to produce a seafloor bottom type map of 9 different bottom types. Examination of sediment sample data by bottom type indicated that each bottom type class had a distinct composition of sediments. Analysis of adult summer flounder, Paralichthys dentatus, and adult silver hake, Merluccius bilinearis, presence/absence data from trawl surveys did not show evidence of strong associations between the species distributions and seafloor bottom type. However, the absence of strong habitat associations may be more attributable to the coarse scale and geographic uncertainty of the trawl sampling data than conclusive evidence that no habitat associations exist for these two species. ?? 2006 Elsevier Ltd. All rights reserved.
Automated Interpretation of Subcellular Patterns in Fluorescence Microscope Images for Location Proteomics

PubMed Central

Chen, Xiang; Velliste, Meel; Murphy, Robert F.

2010-01-01

Proteomics, the large scale identification and characterization of many or all proteins expressed in a given cell type, has become a major area of biological research. In addition to information on protein sequence, structure and expression levels, knowledge of a protein’s subcellular location is essential to a complete understanding of its functions. Currently subcellular location patterns are routinely determined by visual inspection of fluorescence microscope images. We review here research aimed at creating systems for automated, systematic determination of location. These employ numerical feature extraction from images, feature reduction to identify the most useful features, and various supervised learning (classification) and unsupervised learning (clustering) methods. These methods have been shown to perform significantly better than human interpretation of the same images. When coupled with technologies for tagging large numbers of proteins and high-throughput microscope systems, the computational methods reviewed here enable the new subfield of location proteomics. This subfield will make critical contributions in two related areas. First, it will provide structured, high-resolution information on location to enable Systems Biology efforts to simulate cell behavior from the gene level on up. Second, it will provide tools for Cytomics projects aimed at characterizing the behaviors of all cell types before, during and after the onset of various diseases. PMID:16752421
Characterizing the spatial structure of endangered species habitat using geostatistical analysis of IKONOS imagery

USGS Publications Warehouse

Wallace, C.S.A.; Marsh, S.E.

2005-01-01

Our study used geostatistics to extract measures that characterize the spatial structure of vegetated landscapes from satellite imagery for mapping endangered Sonoran pronghorn habitat. Fine spatial resolution IKONOS data provided information at the scale of individual trees or shrubs that permitted analysis of vegetation structure and pattern. We derived images of landscape structure by calculating local estimates of the nugget, sill, and range variogram parameters within 25 ?? 25-m image windows. These variogram parameters, which describe the spatial autocorrelation of the 1-m image pixels, are shown in previous studies to discriminate between different species-specific vegetation associations. We constructed two independent models of pronghorn landscape preference by coupling the derived measures with Sonoran pronghorn sighting data: a distribution-based model and a cluster-based model. The distribution-based model used the descriptive statistics for variogram measures at pronghorn sightings, whereas the cluster-based model used the distribution of pronghorn sightings within clusters of an unsupervised classification of derived images. Both models define similar landscapes, and validation results confirm they effectively predict the locations of an independent set of pronghorn sightings. Such information, although not a substitute for field-based knowledge of the landscape and associated ecological processes, can provide valuable reconnaissance information to guide natural resource management efforts. ?? 2005 Taylor & Francis Group Ltd.
Multi-objective evolutionary optimization for constructing neural networks for virtual reality visual data mining: application to geophysical prospecting.

PubMed

Valdés, Julio J; Barton, Alan J

2007-05-01

A method for the construction of virtual reality spaces for visual data mining using multi-objective optimization with genetic algorithms on nonlinear discriminant (NDA) neural networks is presented. Two neural network layers (the output and the last hidden) are used for the construction of simultaneous solutions for: (i) a supervised classification of data patterns and (ii) an unsupervised similarity structure preservation between the original data matrix and its image in the new space. A set of spaces are constructed from selected solutions along the Pareto front. This strategy represents a conceptual improvement over spaces computed by single-objective optimization. In addition, genetic programming (in particular gene expression programming) is used for finding analytic representations of the complex mappings generating the spaces (a composition of NDA and orthogonal principal components). The presented approach is domain independent and is illustrated via application to the geophysical prospecting of caves.
Development and implementation of a low cost micro computer system for LANDSAT analysis and geographic data base applications

NASA Technical Reports Server (NTRS)

Faust, N.; Jordon, L.

1981-01-01

Since the implementation of the GRID and IMGRID computer programs for multivariate spatial analysis in the early 1970's, geographic data analysis subsequently moved from large computers to minicomputers and now to microcomputers with radical reduction in the costs associated with planning analyses. Programs designed to process LANDSAT data to be used as one element in a geographic data base were used once NIMGRID (new IMGRID), a raster oriented geographic information system, was implemented on the microcomputer. Programs for training field selection, supervised and unsupervised classification, and image enhancement were added. Enhancements to the color graphics capabilities of the microsystem allow display of three channels of LANDSAT data in color infrared format. The basic microcomputer hardware needed to perform NIMGRID and most LANDSAT analyses is listed as well as the software available for LANDSAT processing.
Space-Based Identification of Archaeological Illegal Excavations and a New Automatic Method for Looting Feature Extraction in Desert Areas

NASA Astrophysics Data System (ADS)

Lasaponara, Rosa; Masini, Nicola

2018-06-01

The identification and quantification of disturbance of archaeological sites has been generally approached by visual inspection of optical aerial or satellite pictures. In this paper, we briefly summarize the state of the art of the traditionally satellite-based approaches for looting identification and propose a new automatic method for archaeological looting feature extraction approach (ALFEA). It is based on three steps: the enhancement using spatial autocorrelation, unsupervised classification, and segmentation. ALFEA has been applied to Google Earth images of two test areas, selected in desert environs in Syria (Dura Europos), and in Peru (Cahuachi-Nasca). The reliability of ALFEA was assessed through field surveys in Peru and visual inspection for the Syrian case study. Results from the evaluation procedure showed satisfactory performance from both of the two analysed test cases with a rate of success higher than 90%.
Computational phenotype discovery using unsupervised feature learning over noisy, sparse, and irregular clinical data.

PubMed

Lasko, Thomas A; Denny, Joshua C; Levy, Mia A

2013-01-01

Inferring precise phenotypic patterns from population-scale clinical data is a core computational task in the development of precision, personalized medicine. The traditional approach uses supervised learning, in which an expert designates which patterns to look for (by specifying the learning task and the class labels), and where to look for them (by specifying the input variables). While appropriate for individual tasks, this approach scales poorly and misses the patterns that we don't think to look for. Unsupervised feature learning overcomes these limitations by identifying patterns (or features) that collectively form a compact and expressive representation of the source data, with no need for expert input or labeled examples. Its rising popularity is driven by new deep learning methods, which have produced high-profile successes on difficult standardized problems of object recognition in images. Here we introduce its use for phenotype discovery in clinical data. This use is challenging because the largest source of clinical data - Electronic Medical Records - typically contains noisy, sparse, and irregularly timed observations, rendering them poor substrates for deep learning methods. Our approach couples dirty clinical data to deep learning architecture via longitudinal probability densities inferred using Gaussian process regression. From episodic, longitudinal sequences of serum uric acid measurements in 4368 individuals we produced continuous phenotypic features that suggest multiple population subtypes, and that accurately distinguished (0.97 AUC) the uric-acid signatures of gout vs. acute leukemia despite not being optimized for the task. The unsupervised features were as accurate as gold-standard features engineered by an expert with complete knowledge of the domain, the classification task, and the class labels. Our findings demonstrate the potential for achieving computational phenotype discovery at population scale. We expect such data-driven phenotypes to expose unknown disease variants and subtypes and to provide rich targets for genetic association studies.
Computational Phenotype Discovery Using Unsupervised Feature Learning over Noisy, Sparse, and Irregular Clinical Data

PubMed Central

Lasko, Thomas A.; Denny, Joshua C.; Levy, Mia A.

2013-01-01

Inferring precise phenotypic patterns from population-scale clinical data is a core computational task in the development of precision, personalized medicine. The traditional approach uses supervised learning, in which an expert designates which patterns to look for (by specifying the learning task and the class labels), and where to look for them (by specifying the input variables). While appropriate for individual tasks, this approach scales poorly and misses the patterns that we don’t think to look for. Unsupervised feature learning overcomes these limitations by identifying patterns (or features) that collectively form a compact and expressive representation of the source data, with no need for expert input or labeled examples. Its rising popularity is driven by new deep learning methods, which have produced high-profile successes on difficult standardized problems of object recognition in images. Here we introduce its use for phenotype discovery in clinical data. This use is challenging because the largest source of clinical data – Electronic Medical Records – typically contains noisy, sparse, and irregularly timed observations, rendering them poor substrates for deep learning methods. Our approach couples dirty clinical data to deep learning architecture via longitudinal probability densities inferred using Gaussian process regression. From episodic, longitudinal sequences of serum uric acid measurements in 4368 individuals we produced continuous phenotypic features that suggest multiple population subtypes, and that accurately distinguished (0.97 AUC) the uric-acid signatures of gout vs. acute leukemia despite not being optimized for the task. The unsupervised features were as accurate as gold-standard features engineered by an expert with complete knowledge of the domain, the classification task, and the class labels. Our findings demonstrate the potential for achieving computational phenotype discovery at population scale. We expect such data-driven phenotypes to expose unknown disease variants and subtypes and to provide rich targets for genetic association studies. PMID:23826094
Shakeout: A New Approach to Regularized Deep Neural Network Training.

PubMed

Kang, Guoliang; Li, Jun; Tao, Dacheng

2018-05-01

Recent years have witnessed the success of deep neural networks in dealing with a plenty of practical problems. Dropout has played an essential role in many successful deep neural networks, by inducing regularization in the model training. In this paper, we present a new regularized training approach: Shakeout. Instead of randomly discarding units as Dropout does at the training stage, Shakeout randomly chooses to enhance or reverse each unit's contribution to the next layer. This minor modification of Dropout has the statistical trait: the regularizer induced by Shakeout adaptively combines , and regularization terms. Our classification experiments with representative deep architectures on image datasets MNIST, CIFAR-10 and ImageNet show that Shakeout deals with over-fitting effectively and outperforms Dropout. We empirically demonstrate that Shakeout leads to sparser weights under both unsupervised and supervised settings. Shakeout also leads to the grouping effect of the input units in a layer. Considering the weights in reflecting the importance of connections, Shakeout is superior to Dropout, which is valuable for the deep model compression. Moreover, we demonstrate that Shakeout can effectively reduce the instability of the training process of the deep architecture.
Blood Vessel Extraction in Color Retinal Fundus Images with Enhancement Filtering and Unsupervised Classification

PubMed Central

2017-01-01

Retinal blood vessels have a significant role in the diagnosis and treatment of various retinal diseases such as diabetic retinopathy, glaucoma, arteriosclerosis, and hypertension. For this reason, retinal vasculature extraction is important in order to help specialists for the diagnosis and treatment of systematic diseases. In this paper, a novel approach is developed to extract retinal blood vessel network. Our method comprises four stages: (1) preprocessing stage in order to prepare dataset for segmentation; (2) an enhancement procedure including Gabor, Frangi, and Gauss filters obtained separately before a top-hat transform; (3) a hard and soft clustering stage which includes K-means and Fuzzy C-means (FCM) in order to get binary vessel map; and (4) a postprocessing step which removes falsely segmented isolated regions. The method is tested on color retinal images obtained from STARE and DRIVE databases which are available online. As a result, Gabor filter followed by K-means clustering method achieves 95.94% and 95.71% of accuracy for STARE and DRIVE databases, respectively, which are acceptable for diagnosis systems. PMID:29065611
Machine learning for neuroimaging with scikit-learn.

PubMed

Abraham, Alexandre; Pedregosa, Fabian; Eickenberg, Michael; Gervais, Philippe; Mueller, Andreas; Kossaifi, Jean; Gramfort, Alexandre; Thirion, Bertrand; Varoquaux, Gaël

2014-01-01

Statistical machine learning methods are increasingly used for neuroimaging data analysis. Their main virtue is their ability to model high-dimensional datasets, e.g., multivariate analysis of activation images or resting-state time series. Supervised learning is typically used in decoding or encoding settings to relate brain images to behavioral or clinical observations, while unsupervised learning can uncover hidden structures in sets of images (e.g., resting state functional MRI) or find sub-populations in large cohorts. By considering different functional neuroimaging applications, we illustrate how scikit-learn, a Python machine learning library, can be used to perform some key analysis steps. Scikit-learn contains a very large set of statistical learning algorithms, both supervised and unsupervised, and its application to neuroimaging data provides a versatile tool to study the brain.

Machine learning for neuroimaging with scikit-learn

PubMed Central

Abraham, Alexandre; Pedregosa, Fabian; Eickenberg, Michael; Gervais, Philippe; Mueller, Andreas; Kossaifi, Jean; Gramfort, Alexandre; Thirion, Bertrand; Varoquaux, Gaël

2014-01-01

Statistical machine learning methods are increasingly used for neuroimaging data analysis. Their main virtue is their ability to model high-dimensional datasets, e.g., multivariate analysis of activation images or resting-state time series. Supervised learning is typically used in decoding or encoding settings to relate brain images to behavioral or clinical observations, while unsupervised learning can uncover hidden structures in sets of images (e.g., resting state functional MRI) or find sub-populations in large cohorts. By considering different functional neuroimaging applications, we illustrate how scikit-learn, a Python machine learning library, can be used to perform some key analysis steps. Scikit-learn contains a very large set of statistical learning algorithms, both supervised and unsupervised, and its application to neuroimaging data provides a versatile tool to study the brain. PMID:24600388
Metric Learning for Hyperspectral Image Segmentation

NASA Technical Reports Server (NTRS)

Bue, Brian D.; Thompson, David R.; Gilmore, Martha S.; Castano, Rebecca

2011-01-01

We present a metric learning approach to improve the performance of unsupervised hyperspectral image segmentation. Unsupervised spatial segmentation can assist both user visualization and automatic recognition of surface features. Analysts can use spatially-continuous segments to decrease noise levels and/or localize feature boundaries. However, existing segmentation methods use tasks-agnostic measures of similarity. Here we learn task-specific similarity measures from training data, improving segment fidelity to classes of interest. Multiclass Linear Discriminate Analysis produces a linear transform that optimally separates a labeled set of training classes. The defines a distance metric that generalized to a new scenes, enabling graph-based segmentation that emphasizes key spectral features. We describe tests based on data from the Compact Reconnaissance Imaging Spectrometer (CRISM) in which learned metrics improve segment homogeneity with respect to mineralogical classes.
ECG signal analysis through hidden Markov models.

PubMed

Andreão, Rodrigo V; Dorizzi, Bernadette; Boudy, Jérôme

2006-08-01

This paper presents an original hidden Markov model (HMM) approach for online beat segmentation and classification of electrocardiograms. The HMM framework has been visited because of its ability of beat detection, segmentation and classification, highly suitable to the electrocardiogram (ECG) problem. Our approach addresses a large panel of topics some of them never studied before in other HMM related works: waveforms modeling, multichannel beat segmentation and classification, and unsupervised adaptation to the patient's ECG. The performance was evaluated on the two-channel QT database in terms of waveform segmentation precision, beat detection and classification. Our waveform segmentation results compare favorably to other systems in the literature. We also obtained high beat detection performance with sensitivity of 99.79% and a positive predictivity of 99.96%, using a test set of 59 recordings. Moreover, premature ventricular contraction beats were detected using an original classification strategy. The results obtained validate our approach for real world application.
The effect of the atmosphere on the classification of satellite observations to identify surface features

NASA Technical Reports Server (NTRS)

Fraser, R. S.; Bahethi, O. P.; Al-Abbas, A. H.

1977-01-01

The effect of differences in atmospheric turbidity on the classification of Landsat 1 observations of a rural scene is presented. The observations are classified by an unsupervised clustering technique. These clusters serve as a training set for use of a maximum-likelihood algorithm. The measured radiances in each of the four spectral bands are then changed by amounts measured by Landsat 1. These changes can be associated with a decrease in atmospheric turbidity by a factor of 1.3. The classification of 22% of the pixels changes as a result of the modification. The modified observations are then reclassified as an independent set. Only 3% of the pixels have a different classification than the unmodified set. Hence, if classification errors of rural areas are not to exceed 15%, a new training set has to be developed whenever the difference in turbidity between the training and test sets reaches unity.
Wildlife management by habitat units: A preliminary plan of action

NASA Technical Reports Server (NTRS)

Frentress, C. D.; Frye, R. G.

1975-01-01

Procedures for yielding vegetation type maps were developed using LANDSAT data and a computer assisted classification analysis (LARSYS) to assist in managing populations of wildlife species by defined area units. Ground cover in Travis County, Texas was classified on two occasions using a modified version of the unsupervised approach to classification. The first classification produced a total of 17 classes. Examination revealed that further grouping was justified. A second analysis produced 10 classes which were displayed on printouts which were later color-coded. The final classification was 82 percent accurate. While the classification map appeared to satisfactorily depict the existing vegetation, two classes were determined to contain significant error. The major sources of error could have been eliminated by stratifying cluster sites more closely among previously mapped soil associations that are identified with particular plant associations and by precisely defining class nomenclature using established criteria early in the analysis.
SUSTAIN: a network model of category learning.

PubMed

Love, Bradley C; Medin, Douglas L; Gureckis, Todd M

2004-04-01

SUSTAIN (Supervised and Unsupervised STratified Adaptive Incremental Network) is a model of how humans learn categories from examples. SUSTAIN initially assumes a simple category structure. If simple solutions prove inadequate and SUSTAIN is confronted with a surprising event (e.g., it is told that a bat is a mammal instead of a bird), SUSTAIN recruits an additional cluster to represent the surprising event. Newly recruited clusters are available to explain future events and can themselves evolve into prototypes-attractors-rules. SUSTAIN's discovery of category substructure is affected not only by the structure of the world but by the nature of the learning task and the learner's goals. SUSTAIN successfully extends category learning models to studies of inference learning, unsupervised learning, category construction, and contexts in which identification learning is faster than classification learning.
A novel unsupervised spike sorting algorithm for intracranial EEG.

PubMed

Yadav, R; Shah, A K; Loeb, J A; Swamy, M N S; Agarwal, R

2011-01-01

This paper presents a novel, unsupervised spike classification algorithm for intracranial EEG. The method combines template matching and principal component analysis (PCA) for building a dynamic patient-specific codebook without a priori knowledge of the spike waveforms. The problem of misclassification due to overlapping classes is resolved by identifying similar classes in the codebook using hierarchical clustering. Cluster quality is visually assessed by projecting inter- and intra-clusters onto a 3D plot. Intracranial EEG from 5 patients was utilized to optimize the algorithm. The resulting codebook retains 82.1% of the detected spikes in non-overlapping and disjoint clusters. Initial results suggest a definite role of this method for both rapid review and quantitation of interictal spikes that could enhance both clinical treatment and research studies on epileptic patients.
The cascaded moving k-means and fuzzy c-means clustering algorithms for unsupervised segmentation of malaria images

NASA Astrophysics Data System (ADS)

Abdul-Nasir, Aimi Salihah; Mashor, Mohd Yusoff; Halim, Nurul Hazwani Abd; Mohamed, Zeehaida

2015-05-01

Malaria is a life-threatening parasitic infectious disease that corresponds for nearly one million deaths each year. Due to the requirement of prompt and accurate diagnosis of malaria, the current study has proposed an unsupervised pixel segmentation based on clustering algorithm in order to obtain the fully segmented red blood cells (RBCs) infected with malaria parasites based on the thin blood smear images of P. vivax species. In order to obtain the segmented infected cell, the malaria images are first enhanced by using modified global contrast stretching technique. Then, an unsupervised segmentation technique based on clustering algorithm has been applied on the intensity component of malaria image in order to segment the infected cell from its blood cells background. In this study, cascaded moving k-means (MKM) and fuzzy c-means (FCM) clustering algorithms has been proposed for malaria slide image segmentation. After that, median filter algorithm has been applied to smooth the image as well as to remove any unwanted regions such as small background pixels from the image. Finally, seeded region growing area extraction algorithm has been applied in order to remove large unwanted regions that are still appeared on the image due to their size in which cannot be cleaned by using median filter. The effectiveness of the proposed cascaded MKM and FCM clustering algorithms has been analyzed qualitatively and quantitatively by comparing the proposed cascaded clustering algorithm with MKM and FCM clustering algorithms. Overall, the results indicate that segmentation using the proposed cascaded clustering algorithm has produced the best segmentation performances by achieving acceptable sensitivity as well as high specificity and accuracy values compared to the segmentation results provided by MKM and FCM algorithms.
The New Possibilities from "Big Data" to Overlooked Associations Between Diabetes, Biochemical Parameters, Glucose Control, and Osteoporosis.

PubMed

Kruse, Christian

2018-06-01

To review current practices and technologies within the scope of "Big Data" that can further our understanding of diabetes mellitus and osteoporosis from large volumes of data. "Big Data" techniques involving supervised machine learning, unsupervised machine learning, and deep learning image analysis are presented with examples of current literature. Supervised machine learning can allow us to better predict diabetes-induced osteoporosis and understand relative predictor importance of diabetes-affected bone tissue. Unsupervised machine learning can allow us to understand patterns in data between diabetic pathophysiology and altered bone metabolism. Image analysis using deep learning can allow us to be less dependent on surrogate predictors and use large volumes of images to classify diabetes-induced osteoporosis and predict future outcomes directly from images. "Big Data" techniques herald new possibilities to understand diabetes-induced osteoporosis and ascertain our current ability to classify, understand, and predict this condition.
Unsupervised color image segmentation using a lattice algebra clustering technique

NASA Astrophysics Data System (ADS)

Urcid, Gonzalo; Ritter, Gerhard X.

2011-08-01

In this paper we introduce a lattice algebra clustering technique for segmenting digital images in the Red-Green- Blue (RGB) color space. The proposed technique is a two step procedure. Given an input color image, the first step determines the finite set of its extreme pixel vectors within the color cube by means of the scaled min-W and max-M lattice auto-associative memory matrices, including the minimum and maximum vector bounds. In the second step, maximal rectangular boxes enclosing each extreme color pixel are found using the Chebychev distance between color pixels; afterwards, clustering is performed by assigning each image pixel to its corresponding maximal box. The two steps in our proposed method are completely unsupervised or autonomous. Illustrative examples are provided to demonstrate the color segmentation results including a brief numerical comparison with two other non-maximal variations of the same clustering technique.
Computational efficient unsupervised coastline detection from single-polarization 1-look SAR images of complex coastal environments

NASA Astrophysics Data System (ADS)

Garzelli, Andrea; Zoppetti, Claudia; Pinelli, Gianpaolo

2017-10-01

Coastline detection in synthetic aperture radar (SAR) images is crucial in many application fields, from coastal erosion monitoring to navigation, from damage assessment to security planning for port facilities. The backscattering difference between land and sea is not always documented in SAR imagery, due to the severe speckle noise, especially in 1-look data with high spatial resolution, high sea state, or complex coastal environments. This paper presents an unsupervised, computationally efficient solution to extract the coastline acquired by only one single-polarization 1-look SAR image. Extensive tests on Spotlight COSMO-SkyMed images of complex coastal environments and objective assessment demonstrate the validity of the proposed procedure which is compared to state-of-the-art methods through visual results and with an objective evaluation of the distance between the detected and the true coastline provided by regional authorities.
A Combined Use of Decomposition and Texture for Terrain Classification of Fully Polarimetric SAR Images

NASA Astrophysics Data System (ADS)

Rodionova, N. V.

2007-03-01

This p aper presents two-stag e unsupervised terrain classification of fully polarimetr ic SA R data using Freeman and Durden decomposition based on three simp le scattering mechanisms: surface, volume and double bounce (first step), and textur al features (uncorrelated uniformity , contr ast, inv erse mo men t and entropy) obtained from grey lev el co-occurrence matr ices (GLCM) (second step). Textural f eatures ar e defined in moving w indow 5x5 pixels w ith N=32 (N - number of grey lev els) . This algorith m preserves th e purity of domin ant polarimetric scattering properties and defines textural features in each scatter ing category. It is shown better object discrimin ation after app lying textur e w ith in fix ed scattering category. Speckle r eduction is one of th e main mo ments in imag e interpr etation improvement because of its great influen ce on textur e. Results from unfiltered and Lee filtered polar imetr ic SAR imag es show that the v alues of contrast and en tropy decr ease and th e values of uniformity and inverse moment increase with speckle reduction, that's tru e for all polarizations (HH, VV, HV). Th e d iscr imination b etw een objects increases after speckle f ilter ing. Polar ization influen ce on textur e features is def ined by calculating th e features in SAR images w ith HH , VV and HV polarizations before and after speck le filter ing, and then creating RG B images. It is shown mor e polarization inf luence on textur e features (uniformity , inverse mo ment and entropy) before filtering and less influen ce - after speck le f iltering. I t's not true for contrast wher e polar ization influen ce is not ch anged practically w ith filtering. SIR-C/X-SA R SLC L-band imag es of Moscow r egion are used for illustr ation.
Use of topographic and climatological models in a geographical data base to improve Landsat MSS classification for Olympic National Park

NASA Technical Reports Server (NTRS)

Cibula, William G.; Nyquist, Maurice O.

1987-01-01

An unsupervised computer classification of vegetation/landcover of Olympic National Park and surrounding environs was initially carried out using four bands of Landsat MSS data. The primary objective of the project was to derive a level of landcover classifications useful for park management applications while maintaining an acceptably high level of classification accuracy. Initially, nine generalized vegetation/landcover classes were derived. Overall classification accuracy was 91.7 percent. In an attempt to refine the level of classification, a geographic information system (GIS) approach was employed. Topographic data and watershed boundaries (inferred precipitation/temperature) data were registered with the Landsat MSS data. The resultant boolean operations yielded 21 vegetation/landcover classes while maintaining the same level of classification accuracy. The final classification provided much better identification and location of the major forest types within the park at the same high level of accuracy, and these met the project objective. This classification could now become inputs into a GIS system to help provide answers to park management coupled with other ancillary data programs such as fire management.
Geophysical phenomena classification by artificial neural networks

NASA Technical Reports Server (NTRS)

Gough, M. P.; Bruckner, J. R.

1995-01-01

Space science information systems involve accessing vast data bases. There is a need for an automatic process by which properties of the whole data set can be assimilated and presented to the user. Where data are in the form of spectrograms, phenomena can be detected by pattern recognition techniques. Presented are the first results obtained by applying unsupervised Artificial Neural Networks (ANN's) to the classification of magnetospheric wave spectra. The networks used here were a simple unsupervised Hamming network run on a PC and a more sophisticated CALM network run on a Sparc workstation. The ANN's were compared in their geophysical data recognition performance. CALM networks offer such qualities as fast learning, superiority in generalizing, the ability to continuously adapt to changes in the pattern set, and the possibility to modularize the network to allow the inter-relation between phenomena and data sets. This work is the first step toward an information system interface being developed at Sussex, the Whole Information System Expert (WISE). Phenomena in the data are automatically identified and provided to the user in the form of a data occurrence morphology, the Whole Information System Data Occurrence Morphology (WISDOM), along with relationships to other parameters and phenomena.
A new tool for post-AGB SED classification

NASA Astrophysics Data System (ADS)

Bendjoya, P.; Suarez, O.; Galluccio, L.; Michel, O.

We present the results of an unsupervised classification method applied on a set of 344 spectral energy distributions (SED) of post-AGB stars extracted from the Torun catalogue of Galactic post-AGB stars. This method aims to find a new unbiased method for post-AGB star classification based on the information contained in the IR region of the SED (fluxes, IR excess, colours). We used the data from IRAS and MSX satellites, and from the 2MASS survey. We applied a classification method based on the construction of the dataset of a minimal spanning tree (MST) with the Prim's algorithm. In order to build this tree, different metrics have been tested on both flux and color indices. Our method is able to classify the set of 344 post-AGB stars in 9 distinct groups according to their SEDs.
A Benchmark Dataset and Saliency-guided Stacked Autoencoders for Video-based Salient Object Detection.

PubMed

Li, Jia; Xia, Changqun; Chen, Xiaowu

2017-10-12

Image-based salient object detection (SOD) has been extensively studied in past decades. However, video-based SOD is much less explored due to the lack of large-scale video datasets within which salient objects are unambiguously defined and annotated. Toward this end, this paper proposes a video-based SOD dataset that consists of 200 videos. In constructing the dataset, we manually annotate all objects and regions over 7,650 uniformly sampled keyframes and collect the eye-tracking data of 23 subjects who free-view all videos. From the user data, we find that salient objects in a video can be defined as objects that consistently pop-out throughout the video, and objects with such attributes can be unambiguously annotated by combining manually annotated object/region masks with eye-tracking data of multiple subjects. To the best of our knowledge, it is currently the largest dataset for videobased salient object detection. Based on this dataset, this paper proposes an unsupervised baseline approach for video-based SOD by using saliencyguided stacked autoencoders. In the proposed approach, multiple spatiotemporal saliency cues are first extracted at the pixel, superpixel and object levels. With these saliency cues, stacked autoencoders are constructed in an unsupervised manner that automatically infers a saliency score for each pixel by progressively encoding the high-dimensional saliency cues gathered from the pixel and its spatiotemporal neighbors. In experiments, the proposed unsupervised approach is compared with 31 state-of-the-art models on the proposed dataset and outperforms 30 of them, including 19 imagebased classic (unsupervised or non-deep learning) models, six image-based deep learning models, and five video-based unsupervised models. Moreover, benchmarking results show that the proposed dataset is very challenging and has the potential to boost the development of video-based SOD.
Unsupervised image matching based on manifold alignment.

PubMed

Pei, Yuru; Huang, Fengchun; Shi, Fuhao; Zha, Hongbin

2012-08-01

This paper challenges the issue of automatic matching between two image sets with similar intrinsic structures and different appearances, especially when there is no prior correspondence. An unsupervised manifold alignment framework is proposed to establish correspondence between data sets by a mapping function in the mutual embedding space. We introduce a local similarity metric based on parameterized distance curves to represent the connection of one point with the rest of the manifold. A small set of valid feature pairs can be found without manual interactions by matching the distance curve of one manifold with the curve cluster of the other manifold. To avoid potential confusions in image matching, we propose an extended affine transformation to solve the nonrigid alignment in the embedding space. The comparatively tight alignments and the structure preservation can be obtained simultaneously. The point pairs with the minimum distance after alignment are viewed as the matchings. We apply manifold alignment to image set matching problems. The correspondence between image sets of different poses, illuminations, and identities can be established effectively by our approach.
Model–Free Visualization of Suspicious Lesions in Breast MRI Based on Supervised and Unsupervised Learning

PubMed Central

Twellmann, Thorsten; Meyer-Baese, Anke; Lange, Oliver; Foo, Simon; Nattkemper, Tim W.

2008-01-01

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) has become an important tool in breast cancer diagnosis, but evaluation of multitemporal 3D image data holds new challenges for human observers. To aid the image analysis process, we apply supervised and unsupervised pattern recognition techniques for computing enhanced visualizations of suspicious lesions in breast MRI data. These techniques represent an important component of future sophisticated computer-aided diagnosis (CAD) systems and support the visual exploration of spatial and temporal features of DCE-MRI data stemming from patients with confirmed lesion diagnosis. By taking into account the heterogeneity of cancerous tissue, these techniques reveal signals with malignant, benign and normal kinetics. They also provide a regional subclassification of pathological breast tissue, which is the basis for pseudo-color presentations of the image data. Intelligent medical systems are expected to have substantial implications in healthcare politics by contributing to the diagnosis of indeterminate breast lesions by non-invasive imaging. PMID:19255616
CNN universal machine as classificaton platform: an art-like clustering algorithm.

PubMed

Bálya, David

2003-12-01

Fast and robust classification of feature vectors is a crucial task in a number of real-time systems. A cellular neural/nonlinear network universal machine (CNN-UM) can be very efficient as a feature detector. The next step is to post-process the results for object recognition. This paper shows how a robust classification scheme based on adaptive resonance theory (ART) can be mapped to the CNN-UM. Moreover, this mapping is general enough to include different types of feed-forward neural networks. The designed analogic CNN algorithm is capable of classifying the extracted feature vectors keeping the advantages of the ART networks, such as robust, plastic and fault-tolerant behaviors. An analogic algorithm is presented for unsupervised classification with tunable sensitivity and automatic new class creation. The algorithm is extended for supervised classification. The presented binary feature vector classification is implemented on the existing standard CNN-UM chips for fast classification. The experimental evaluation shows promising performance after 100% accuracy on the training set.
A Deep Convolutional Coupling Network for Change Detection Based on Heterogeneous Optical and Radar Images.

PubMed

Liu, Jia; Gong, Maoguo; Qin, Kai; Zhang, Puzhao

2018-03-01

We propose an unsupervised deep convolutional coupling network for change detection based on two heterogeneous images acquired by optical sensors and radars on different dates. Most existing change detection methods are based on homogeneous images. Due to the complementary properties of optical and radar sensors, there is an increasing interest in change detection based on heterogeneous images. The proposed network is symmetric with each side consisting of one convolutional layer and several coupling layers. The two input images connected with the two sides of the network, respectively, are transformed into a feature space where their feature representations become more consistent. In this feature space, the different map is calculated, which then leads to the ultimate detection map by applying a thresholding algorithm. The network parameters are learned by optimizing a coupling function. The learning process is unsupervised, which is different from most existing change detection methods based on heterogeneous images. Experimental results on both homogenous and heterogeneous images demonstrate the promising performance of the proposed network compared with several existing approaches.

Wavelet-based unsupervised learning method for electrocardiogram suppression in surface electromyograms.

PubMed

Niegowski, Maciej; Zivanovic, Miroslav

2016-03-01

We present a novel approach aimed at removing electrocardiogram (ECG) perturbation from single-channel surface electromyogram (EMG) recordings by means of unsupervised learning of wavelet-based intensity images. The general idea is to combine the suitability of certain wavelet decomposition bases which provide sparse electrocardiogram time-frequency representations, with the capacity of non-negative matrix factorization (NMF) for extracting patterns from images. In order to overcome convergence problems which often arise in NMF-related applications, we design a novel robust initialization strategy which ensures proper signal decomposition in a wide range of ECG contamination levels. Moreover, the method can be readily used because no a priori knowledge or parameter adjustment is needed. The proposed method was evaluated on real surface EMG signals against two state-of-the-art unsupervised learning algorithms and a singular spectrum analysis based method. The results, expressed in terms of high-to-low energy ratio, normalized median frequency, spectral power difference and normalized average rectified value, suggest that the proposed method enables better ECG-EMG separation quality than the reference methods. Copyright © 2015 IPEM. Published by Elsevier Ltd. All rights reserved.
Galaxy morphology - An unsupervised machine learning approach

NASA Astrophysics Data System (ADS)

Schutter, A.; Shamir, L.

2015-09-01

Structural properties poses valuable information about the formation and evolution of galaxies, and are important for understanding the past, present, and future universe. Here we use unsupervised machine learning methodology to analyze a network of similarities between galaxy morphological types, and automatically deduce a morphological sequence of galaxies. Application of the method to the EFIGI catalog show that the morphological scheme produced by the algorithm is largely in agreement with the De Vaucouleurs system, demonstrating the ability of computer vision and machine learning methods to automatically profile galaxy morphological sequences. The unsupervised analysis method is based on comprehensive computer vision techniques that compute the visual similarities between the different morphological types. Rather than relying on human cognition, the proposed system deduces the similarities between sets of galaxy images in an automatic manner, and is therefore not limited by the number of galaxies being analyzed. The source code of the method is publicly available, and the protocol of the experiment is included in the paper so that the experiment can be replicated, and the method can be used to analyze user-defined datasets of galaxy images.
Validation of a free software for unsupervised assessment of abdominal fat in MRI.

PubMed

Maddalo, Michele; Zorza, Ivan; Zubani, Stefano; Nocivelli, Giorgio; Calandra, Giulio; Soldini, Pierantonio; Mascaro, Lorella; Maroldi, Roberto

2017-05-01

To demonstrate the accuracy of an unsupervised (fully automated) software for fat segmentation in magnetic resonance imaging. The proposed software is a freeware solution developed in ImageJ that enables the quantification of metabolically different adipose tissues in large cohort studies. The lumbar part of the abdomen (19cm in craniocaudal direction, centered in L3) of eleven healthy volunteers (age range: 21-46years, BMI range: 21.7-31.6kg/m 2 ) was examined in a breath hold on expiration with a GE T1 Dixon sequence. Single-slice and volumetric data were considered for each subject. The results of the visceral and subcutaneous adipose tissue assessments obtained by the unsupervised software were compared to supervised segmentations of reference. The associated statistical analysis included Pearson correlations, Bland-Altman plots and volumetric differences (VD % ). Values calculated by the unsupervised software significantly correlated with corresponding supervised segmentations of reference for both subcutaneous adipose tissue - SAT (R=0.9996, p<0.001) and visceral adipose tissue - VAT (R=0.995, p<0.001). Bland-Altman plots showed the absence of systematic errors and a limited spread of the differences. In the single-slice analysis, VD % were (1.6±2.9)% for SAT and (4.9±6.9)% for VAT. In the volumetric analysis, VD % were (1.3±0.9)% for SAT and (2.9±2.7)% for VAT. The developed software is capable of segmenting the metabolically different adipose tissues with a high degree of accuracy. This free add-on software for ImageJ can easily have a widespread and enable large-scale population studies regarding the adipose tissue and its related diseases. Copyright © 2017 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Extending bicluster analysis to annotate unclassified ORFs and predict novel functional modules using expression data

PubMed Central

Bryan, Kenneth; Cunningham, Pádraig

2008-01-01

Background Microarrays have the capacity to measure the expressions of thousands of genes in parallel over many experimental samples. The unsupervised classification technique of bicluster analysis has been employed previously to uncover gene expression correlations over subsets of samples with the aim of providing a more accurate model of the natural gene functional classes. This approach also has the potential to aid functional annotation of unclassified open reading frames (ORFs). Until now this aspect of biclustering has been under-explored. In this work we illustrate how bicluster analysis may be extended into a 'semi-supervised' ORF annotation approach referred to as BALBOA. Results The efficacy of the BALBOA ORF classification technique is first assessed via cross validation and compared to a multi-class k-Nearest Neighbour (kNN) benchmark across three independent gene expression datasets. BALBOA is then used to assign putative functional annotations to unclassified yeast ORFs. These predictions are evaluated using existing experimental and protein sequence information. Lastly, we employ a related semi-supervised method to predict the presence of novel functional modules within yeast. Conclusion In this paper we demonstrate how unsupervised classification methods, such as bicluster analysis, may be extended using of available annotations to form semi-supervised approaches within the gene expression analysis domain. We show that such methods have the potential to improve upon supervised approaches and shed new light on the functions of unclassified ORFs and their co-regulation. PMID:18831786
AUTOMATED UNSUPERVISED CLASSIFICATION OF THE SLOAN DIGITAL SKY SURVEY STELLAR SPECTRA USING k-MEANS CLUSTERING

DOE Office of Scientific and Technical Information (OSTI.GOV)

Sanchez Almeida, J.; Allende Prieto, C., E-mail: jos@iac.es, E-mail: callende@iac.es

2013-01-20

Large spectroscopic surveys require automated methods of analysis. This paper explores the use of k-means clustering as a tool for automated unsupervised classification of massive stellar spectral catalogs. The classification criteria are defined by the data and the algorithm, with no prior physical framework. We work with a representative set of stellar spectra associated with the Sloan Digital Sky Survey (SDSS) SEGUE and SEGUE-2 programs, which consists of 173,390 spectra from 3800 to 9200 A sampled on 3849 wavelengths. We classify the original spectra as well as the spectra with the continuum removed. The second set only contains spectral lines,more » and it is less dependent on uncertainties of the flux calibration. The classification of the spectra with continuum renders 16 major classes. Roughly speaking, stars are split according to their colors, with enough finesse to distinguish dwarfs from giants of the same effective temperature, but with difficulties to separate stars with different metallicities. There are classes corresponding to particular MK types, intrinsically blue stars, dust-reddened, stellar systems, and also classes collecting faulty spectra. Overall, there is no one-to-one correspondence between the classes we derive and the MK types. The classification of spectra without continuum renders 13 classes, the color separation is not so sharp, but it distinguishes stars of the same effective temperature and different metallicities. Some classes thus obtained present a fairly small range of physical parameters (200 K in effective temperature, 0.25 dex in surface gravity, and 0.35 dex in metallicity), so that the classification can be used to estimate the main physical parameters of some stars at a minimum computational cost. We also analyze the outliers of the classification. Most of them turn out to be failures of the reduction pipeline, but there are also high redshift QSOs, multiple stellar systems, dust-reddened stars, galaxies, and, finally, odd spectra whose nature we have not deciphered. The template spectra representative of the classes are publicly available in the online journal.« less
Sequence-structure relationship study in all-α transmembrane proteins using an unsupervised learning approach.

PubMed

Esque, Jérémy; Urbain, Aurélie; Etchebest, Catherine; de Brevern, Alexandre G

2015-11-01

Transmembrane proteins (TMPs) are major drug targets, but the knowledge of their precise topology structure remains highly limited compared with globular proteins. In spite of the difficulties in obtaining their structures, an important effort has been made these last years to increase their number from an experimental and computational point of view. In view of this emerging challenge, the development of computational methods to extract knowledge from these data is crucial for the better understanding of their functions and in improving the quality of structural models. Here, we revisit an efficient unsupervised learning procedure, called Hybrid Protein Model (HPM), which is applied to the analysis of transmembrane proteins belonging to the all-α structural class. HPM method is an original classification procedure that efficiently combines sequence and structure learning. The procedure was initially applied to the analysis of globular proteins. In the present case, HPM classifies a set of overlapping protein fragments, extracted from a non-redundant databank of TMP 3D structure. After fine-tuning of the learning parameters, the optimal classification results in 65 clusters. They represent at best similar relationships between sequence and local structure properties of TMPs. Interestingly, HPM distinguishes among the resulting clusters two helical regions with distinct hydrophobic patterns. This underlines the complexity of the topology of these proteins. The HPM classification enlightens unusual relationship between amino acids in TMP fragments, which can be useful to elaborate new amino acids substitution matrices. Finally, two challenging applications are described: the first one aims at annotating protein functions (channel or not), the second one intends to assess the quality of the structures (X-ray or models) via a new scoring function deduced from the HPM classification.
Hydrometeor classification through statistical clustering of polarimetric radar measurements: a semi-supervised approach

NASA Astrophysics Data System (ADS)

Besic, Nikola; Ventura, Jordi Figueras i.; Grazioli, Jacopo; Gabella, Marco; Germann, Urs; Berne, Alexis

2016-09-01

Polarimetric radar-based hydrometeor classification is the procedure of identifying different types of hydrometeors by exploiting polarimetric radar observations. The main drawback of the existing supervised classification methods, mostly based on fuzzy logic, is a significant dependency on a presumed electromagnetic behaviour of different hydrometeor types. Namely, the results of the classification largely rely upon the quality of scattering simulations. When it comes to the unsupervised approach, it lacks the constraints related to the hydrometeor microphysics. The idea of the proposed method is to compensate for these drawbacks by combining the two approaches in a way that microphysical hypotheses can, to a degree, adjust the content of the classes obtained statistically from the observations. This is done by means of an iterative approach, performed offline, which, in a statistical framework, examines clustered representative polarimetric observations by comparing them to the presumed polarimetric properties of each hydrometeor class. Aside from comparing, a routine alters the content of clusters by encouraging further statistical clustering in case of non-identification. By merging all identified clusters, the multi-dimensional polarimetric signatures of various hydrometeor types are obtained for each of the studied representative datasets, i.e. for each radar system of interest. These are depicted by sets of centroids which are then employed in operational labelling of different hydrometeors. The method has been applied on three C-band datasets, each acquired by different operational radar from the MeteoSwiss Rad4Alp network, as well as on two X-band datasets acquired by two research mobile radars. The results are discussed through a comparative analysis which includes a corresponding supervised and unsupervised approach, emphasising the operational potential of the proposed method.
Linear time relational prototype based learning.

PubMed

Gisbrecht, Andrej; Mokbel, Bassam; Schleif, Frank-Michael; Zhu, Xibin; Hammer, Barbara

2012-10-01

Prototype based learning offers an intuitive interface to inspect large quantities of electronic data in supervised or unsupervised settings. Recently, many techniques have been extended to data described by general dissimilarities rather than Euclidean vectors, so-called relational data settings. Unlike the Euclidean counterparts, the techniques have quadratic time complexity due to the underlying quadratic dissimilarity matrix. Thus, they are infeasible already for medium sized data sets. The contribution of this article is twofold: On the one hand we propose a novel supervised prototype based classification technique for dissimilarity data based on popular learning vector quantization (LVQ), on the other hand we transfer a linear time approximation technique, the Nyström approximation, to this algorithm and an unsupervised counterpart, the relational generative topographic mapping (GTM). This way, linear time and space methods result. We evaluate the techniques on three examples from the biomedical domain.
Colour image segmentation using unsupervised clustering technique for acute leukemia images

NASA Astrophysics Data System (ADS)

Halim, N. H. Abd; Mashor, M. Y.; Nasir, A. S. Abdul; Mustafa, N.; Hassan, R.

2015-05-01

Colour image segmentation has becoming more popular for computer vision due to its important process in most medical analysis tasks. This paper proposes comparison between different colour components of RGB(red, green, blue) and HSI (hue, saturation, intensity) colour models that will be used in order to segment the acute leukemia images. First, partial contrast stretching is applied on leukemia images to increase the visual aspect of the blast cells. Then, an unsupervised moving k-means clustering algorithm is applied on the various colour components of RGB and HSI colour models for the purpose of segmentation of blast cells from the red blood cells and background regions in leukemia image. Different colour components of RGB and HSI colour models have been analyzed in order to identify the colour component that can give the good segmentation performance. The segmented images are then processed using median filter and region growing technique to reduce noise and smooth the images. The results show that segmentation using saturation component of HSI colour model has proven to be the best in segmenting nucleus of the blast cells in acute leukemia image as compared to the other colour components of RGB and HSI colour models.
Investigation using data in Alabama from ERTS-A

NASA Technical Reports Server (NTRS)

Henry, H. R. (Principal Investigator)

1972-01-01

There are no author-identified significant results in this report. Brief summaries are presented of accomplishments by the state of Alabama in the areas of: (1) investigation of environmental factors; (2) land use compilation; (3) data processing for land use compilation; (4) photo-reproduction and unsupervised land use classification from digital tape; (5) data collection buoys; and (6) activities of the Geological Survey of Alabama.
An Integrated approach to the Space Situational Awareness Problem

DTIC Science & Technology

2016-12-15

data coming from the sensors. We developed particle-based Gaussian Mixture Filters that are immune to the “curse of dimensionality”/ “particle...depletion” problem inherent in particle filtering . This method maps the data assimilation/ filtering problem into an unsupervised learning problem. Results...Gaussian Mixture Filters ; particle depletion; Finite Set Statistics 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT UU 18. NUMBER OF PAGES 1
Simultaneously Discovering and Localizing Common Objects in Wild Images.

PubMed

Wang, Zhenzhen; Yuan, Junsong

2018-09-01

Motivated by the recent success of supervised and weakly supervised common object discovery, in this paper, we move forward one step further to tackle common object discovery in a fully unsupervised way. Generally, object co-localization aims at simultaneously localizing objects of the same class across a group of images. Traditional object localization/detection usually trains specific object detectors which require bounding box annotations of object instances, or at least image-level labels to indicate the presence/absence of objects in an image. Given a collection of images without any annotations, our proposed fully unsupervised method is to simultaneously discover images that contain common objects and also localize common objects in corresponding images. Without requiring to know the total number of common objects, we formulate this unsupervised object discovery as a sub-graph mining problem from a weighted graph of object proposals, where nodes correspond to object proposals, and edges represent the similarities between neighbouring proposals. The positive images and common objects are jointly discovered by finding sub-graphs of strongly connected nodes, with each sub-graph capturing one object pattern. The optimization problem can be efficiently solved by our proposed maximal-flow-based algorithm. Instead of assuming that each image contains only one common object, our proposed solution can better address wild images where each image may contain multiple common objects or even no common object. Moreover, our proposed method can be easily tailored to the task of image retrieval in which the nodes correspond to the similarity between query and reference images. Extensive experiments on PASCAL VOC 2007 and Object Discovery data sets demonstrate that even without any supervision, our approach can discover/localize common objects of various classes in the presence of scale, view point, appearance variation, and partial occlusions. We also conduct broad experiments on image retrieval benchmarks, Holidays and Oxford5k data sets, to show that our proposed method, which considers both the similarity between query and reference images and also similarities among reference images, can help to improve the retrieval results significantly.
Unsupervised learning toward brain imaging data analysis: cigarette craving and resistance related neuronal activations from functional magnetic resonance imaging data analysis

NASA Astrophysics Data System (ADS)

Kim, Dong-Youl; Lee, Jong-Hwan

2014-05-01

A data-driven unsupervised learning such as an independent component analysis was gainfully applied to bloodoxygenation- level-dependent (BOLD) functional magnetic resonance imaging (fMRI) data compared to a model-based general linear model (GLM). This is due to an ability of this unsupervised learning method to extract a meaningful neuronal activity from BOLD signal that is a mixture of confounding non-neuronal artifacts such as head motions and physiological artifacts as well as neuronal signals. In this study, we support this claim by identifying neuronal underpinnings of cigarette craving and cigarette resistance. The fMRI data were acquired from heavy cigarette smokers (n = 14) while they alternatively watched images with and without cigarette smoking. During acquisition of two fMRI runs, they were asked to crave when they watched cigarette smoking images or to resist the urge to smoke. Data driven approaches of group independent component analysis (GICA) method based on temporal concatenation (TC) and TCGICA with an extension of iterative dual-regression (TC-GICA-iDR) were applied to the data. From the results, cigarette craving and cigarette resistance related neuronal activations were identified in the visual area and superior frontal areas, respectively with a greater statistical significance from the TC-GICA-iDR method than the TC-GICA method. On the other hand, the neuronal activity levels in many of these regions were not statistically different from the GLM method between the cigarette craving and cigarette resistance due to potentially aberrant BOLD signals.
Simple adaptive sparse representation based classification schemes for EEG based brain-computer interface applications.

PubMed

Shin, Younghak; Lee, Seungchan; Ahn, Minkyu; Cho, Hohyun; Jun, Sung Chan; Lee, Heung-No

2015-11-01

One of the main problems related to electroencephalogram (EEG) based brain-computer interface (BCI) systems is the non-stationarity of the underlying EEG signals. This results in the deterioration of the classification performance during experimental sessions. Therefore, adaptive classification techniques are required for EEG based BCI applications. In this paper, we propose simple adaptive sparse representation based classification (SRC) schemes. Supervised and unsupervised dictionary update techniques for new test data and a dictionary modification method by using the incoherence measure of the training data are investigated. The proposed methods are very simple and additional computation for the re-training of the classifier is not needed. The proposed adaptive SRC schemes are evaluated using two BCI experimental datasets. The proposed methods are assessed by comparing classification results with the conventional SRC and other adaptive classification methods. On the basis of the results, we find that the proposed adaptive schemes show relatively improved classification accuracy as compared to conventional methods without requiring additional computation. Copyright © 2015 Elsevier Ltd. All rights reserved.
Translating statistical images to text summaries for partially sighted persons on mobile devices: iconic image maps approach

NASA Astrophysics Data System (ADS)

Williams, Godfried B.

2005-03-01

This paper attempts to demonstrate a novel based idea for transforming statistical image data to text using autoassociative and unsupervised artificial neural network and iconic image maps using the shape and texture genetic algorithm, underlying concepts translating the image data to text. Full details of experiments could be assessed at http://www.uel.ac.uk/seis/applications/.
Multiresource analysis and information system concepts for incorporating LANDSAT and GIS technology into large area forest surveys. [South Carolina

NASA Technical Reports Server (NTRS)

Langley, P. G.

1981-01-01

A method of relating different classifications at each stage of a multistage, multiresource inventory using remotely sensed imagery is discussed. A class transformation matrix allowing the conversion of a set of proportions at one stage, to a set of proportions at the subsequent stage through use of a linear model, is described. The technique was tested by applying it to Kershaw County, South Carolina. Unsupervised LANDSAT spectral classifications were correlated with interpretations of land use aerial photography, the correlations employed to estimate land use classifications using the linear model, and the land use proportions used to stratify current annual increment (CAI) field plot data to obtain a total CAI for the county. The estimate differed by 1% from the published figure for land use. Potential sediment loss and a variety of land use classifications were also obtained.
A graph lattice approach to maintaining and learning dense collections of subgraphs as image features.

PubMed

Saund, Eric

2013-10-01

Effective object and scene classification and indexing depend on extraction of informative image features. This paper shows how large families of complex image features in the form of subgraphs can be built out of simpler ones through construction of a graph lattice—a hierarchy of related subgraphs linked in a lattice. Robustness is achieved by matching many overlapping and redundant subgraphs, which allows the use of inexpensive exact graph matching, instead of relying on expensive error-tolerant graph matching to a minimal set of ideal model graphs. Efficiency in exact matching is gained by exploitation of the graph lattice data structure. Additionally, the graph lattice enables methods for adaptively growing a feature space of subgraphs tailored to observed data. We develop the approach in the domain of rectilinear line art, specifically for the practical problem of document forms recognition. We are especially interested in methods that require only one or very few labeled training examples per category. We demonstrate two approaches to using the subgraph features for this purpose. Using a bag-of-words feature vector we achieve essentially single-instance learning on a benchmark forms database, following an unsupervised clustering stage. Further performance gains are achieved on a more difficult dataset using a feature voting method and feature selection procedure.
Integrated change detection and temporal trajectory analysis of coastal wetlands using high spatial resolution Korean Multi-Purpose Satellite series imagery

NASA Astrophysics Data System (ADS)

Nguyen, Hoang Hai; Tran, Hien; Sunwoo, Wooyeon; Yi, Jong-hyuk; Kim, Dongkyun; Choi, Minha

2017-04-01

A series of multispectral high-resolution Korean Multi-Purpose Satellite (KOMPSAT) images was used to detect the geographical changes in four different tidal flats between the Yellow Sea and the west coast of South Korea. The method of unsupervised classification was used to generate a series of land use/land cover (LULC) maps from satellite images, which were then used as input for temporal trajectory analysis to detect the temporal change of coastal wetlands and its association with natural and anthropogenic activities. The accurately classified LULC maps of KOMPSAT images, with overall accuracy ranging from 83.34% to 95.43%, indicate that these multispectral high-resolution satellite data are highly applicable to the generation of high-quality thematic maps for extracting wetlands. The result of the trajectory analysis showed that, while the variation of the tidal flats in the Gyeonggi and Jeollabuk provinces was well correlated with the regular tidal regimes, the reductive trajectory of the wetland areas belonging to the Saemangeum province was caused by a high degree of human-induced activities including large reclamation and urbanization. The conservation of the Jeungdo Wetland Protected Area in the Jeollanam province revealed that effective social and environmental policies could help in protecting coastal wetlands from degradation.
Monitoring wetlands change using LANDSAT data

NASA Technical Reports Server (NTRS)

Hardin, D. L.

1981-01-01

A wetlands monitoring study was initiated as part of Delaware's LANDSAT applications demonstration project. Classifications of digital data are conducted in an effort to determine the location and acreage of wetlands loss or gain, species conversion, and application for the inventory and typing of freshwater wetlands. A multi-seasonal approach is employed to compare data from two different years. Unsupervised classifications were conducted for two of the four dates examined. Initial results indicate the multi-seasonal approach allows much better separation of wetland types for both tidal and non-tidal wetlands than either season alone. Change detection is possible but generally misses the small acreages now impacted by man.
myBlackBox: Blackbox Mobile Cloud Systems for Personalized Unusual Event Detection.

PubMed

Ahn, Junho; Han, Richard

2016-05-23

We demonstrate the feasibility of constructing a novel and practical real-world mobile cloud system, called myBlackBox, that efficiently fuses multimodal smartphone sensor data to identify and log unusual personal events in mobile users' daily lives. The system incorporates a hybrid architectural design that combines unsupervised classification of audio, accelerometer and location data with supervised joint fusion classification to achieve high accuracy, customization, convenience and scalability. We show the feasibility of myBlackBox by implementing and evaluating this end-to-end system that combines Android smartphones with cloud servers, deployed for 15 users over a one-month period.

myBlackBox: Blackbox Mobile Cloud Systems for Personalized Unusual Event Detection

PubMed Central

Ahn, Junho; Han, Richard

2016-01-01

We demonstrate the feasibility of constructing a novel and practical real-world mobile cloud system, called myBlackBox, that efficiently fuses multimodal smartphone sensor data to identify and log unusual personal events in mobile users’ daily lives. The system incorporates a hybrid architectural design that combines unsupervised classification of audio, accelerometer and location data with supervised joint fusion classification to achieve high accuracy, customization, convenience and scalability. We show the feasibility of myBlackBox by implementing and evaluating this end-to-end system that combines Android smartphones with cloud servers, deployed for 15 users over a one-month period. PMID:27223292
Mixtures of GAMs for habitat suitability analysis with overdispersed presence / absence data

PubMed Central

Pleydell, David R.J.; Chrétien, Stéphane

2009-01-01

A new approach to species distribution modelling based on unsupervised classification via a finite mixture of GAMs incorporating habitat suitability curves is proposed. A tailored EM algorithm is outlined for computing maximum likelihood estimates. Several submodels incorporating various parameter constraints are explored. Simulation studies confirm, that under certain constraints, the habitat suitability curves are recovered with good precision. The method is also applied to a set of real data concerning presence/absence of observable small mammal indices collected on the Tibetan plateau. The resulting classification was found to correspond to species-level differences in habitat preference described in previous ecological work. PMID:20401331
Eigenvalue-eigenvector decomposition (EED) analysis of dissimilarity and covariance matrix obtained from total synchronous fluorescence spectral (TSFS) data sets of herbal preparations: Optimizing the classification approach

NASA Astrophysics Data System (ADS)

Tarai, Madhumita; Kumar, Keshav; Divya, O.; Bairi, Partha; Mishra, Kishor Kumar; Mishra, Ashok Kumar

2017-09-01

The present work compares the dissimilarity and covariance based unsupervised chemometric classification approaches by taking the total synchronous fluorescence spectroscopy data sets acquired for the cumin and non-cumin based herbal preparations. The conventional decomposition method involves eigenvalue-eigenvector analysis of the covariance of the data set and finds the factors that can explain the overall major sources of variation present in the data set. The conventional approach does this irrespective of the fact that the samples belong to intrinsically different groups and hence leads to poor class separation. The present work shows that classification of such samples can be optimized by performing the eigenvalue-eigenvector decomposition on the pair-wise dissimilarity matrix.
Commodity cluster and hardware-based massively parallel implementations of hyperspectral imaging algorithms

NASA Astrophysics Data System (ADS)

Plaza, Antonio; Chang, Chein-I.; Plaza, Javier; Valencia, David

2006-05-01

The incorporation of hyperspectral sensors aboard airborne/satellite platforms is currently producing a nearly continual stream of multidimensional image data, and this high data volume has soon introduced new processing challenges. The price paid for the wealth spatial and spectral information available from hyperspectral sensors is the enormous amounts of data that they generate. Several applications exist, however, where having the desired information calculated quickly enough for practical use is highly desirable. High computing performance of algorithm analysis is particularly important in homeland defense and security applications, in which swift decisions often involve detection of (sub-pixel) military targets (including hostile weaponry, camouflage, concealment, and decoys) or chemical/biological agents. In order to speed-up computational performance of hyperspectral imaging algorithms, this paper develops several fast parallel data processing techniques. Techniques include four classes of algorithms: (1) unsupervised classification, (2) spectral unmixing, and (3) automatic target recognition, and (4) onboard data compression. A massively parallel Beowulf cluster (Thunderhead) at NASA's Goddard Space Flight Center in Maryland is used to measure parallel performance of the proposed algorithms. In order to explore the viability of developing onboard, real-time hyperspectral data compression algorithms, a Xilinx Virtex-II field programmable gate array (FPGA) is also used in experiments. Our quantitative and comparative assessment of parallel techniques and strategies may help image analysts in selection of parallel hyperspectral algorithms for specific applications.
A neuromorphic approach to satellite image understanding

NASA Astrophysics Data System (ADS)

Partsinevelos, Panagiotis; Perakakis, Manolis

2014-05-01

Remote sensing satellite imagery provides high altitude, top viewing aspects of large geographic regions and as such the depicted features are not always easily recognizable. Nevertheless, geoscientists familiar to remote sensing data, gradually gain experience and enhance their satellite image interpretation skills. The aim of this study is to devise a novel computational neuro-centered classification approach for feature extraction and image understanding. Object recognition through image processing practices is related to a series of known image/feature based attributes including size, shape, association, texture, etc. The objective of the study is to weight these attribute values towards the enhancement of feature recognition. The key cognitive experimentation concern is to define the point when a user recognizes a feature as it varies in terms of the above mentioned attributes and relate it with their corresponding values. Towards this end, we have set up an experimentation methodology that utilizes cognitive data from brain signals (EEG) and eye gaze data (eye tracking) of subjects watching satellite images of varying attributes; this allows the collection of rich real-time data that will be used for designing the image classifier. Since the data are already labeled by users (using an input device) a first step is to compare the performance of various machine-learning algorithms on the collected data. On the long-run, the aim of this work would be to investigate the automatic classification of unlabeled images (unsupervised learning) based purely on image attributes. The outcome of this innovative process is twofold: First, in an abundance of remote sensing image datasets we may define the essential image specifications in order to collect the appropriate data for each application and improve processing and resource efficiency. E.g. for a fault extraction application in a given scale a medium resolution 4-band image, may be more effective than costly, multispectral, very high resolution imagery. Second, we attempt to relate the experienced against the non-experienced user understanding in order to indirectly assess the possible limits of purely computational systems. In other words, obtain the conceptual limits of computation vs human cognition concerning feature recognition from satellite imagery. Preliminary results of this pilot study show relations between collected data and differentiation of the image attributes which indicates that our methodology can lead to important results.
Creating high-resolution time series land-cover classifications in rapidly changing forested areas with BULC-U in Google Earth Engine

NASA Astrophysics Data System (ADS)

Cardille, J. A.; Lee, J.

2017-12-01

With the opening of the Landsat archive, there is a dramatically increased potential for creating high-quality time series of land use/land-cover (LULC) classifications derived from remote sensing. Although LULC time series are appealing, their creation is typically challenging in two fundamental ways. First, there is a need to create maximally correct LULC maps for consideration at each time step; and second, there is a need to have the elements of the time series be consistent with each other, without pixels that flip improbably between covers due only to unavoidable, stray classification errors. We have developed the Bayesian Updating of Land Cover - Unsupervised (BULC-U) algorithm to address these challenges simultaneously, and introduce and apply it here for two related but distinct purposes. First, with minimal human intervention, we produced an internally consistent, high-accuracy LULC time series in rapidly changing Mato Grosso, Brazil for a time interval (1986-2000) in which cropland area more than doubled. The spatial and temporal resolution of the 59 LULC snapshots allows users to witness the establishment of towns and farms at the expense of forest. The new time series could be used by policy-makers and analysts to unravel important considerations for conservation and management, including the timing and location of past development, the rate and nature of changes in forest connectivity, the connection with road infrastructure, and more. The second application of BULC-U is to sharpen the well-known GlobCover 2009 classification from 300m to 30m, while improving accuracy measures for every class. The greatly improved resolution and accuracy permits a better representation of the true LULC proportions, the use of this map in models, and quantification of the potential impacts of changes. Given that there may easily be thousands and potentially millions of images available to harvest for an LULC time series, it is imperative to build useful algorithms requiring minimal human intervention. Through image segmentation and classification, BULC-U allows us to use both the spectral and spatial characteristics of imagery to sharpen classifications and create time series. It is hoped that this study may allow us and other users of this new method to consider time series across ever larger areas.
Information-Based Approach to Unsupervised Machine Learning

DTIC Science & Technology

2013-06-19

Leibler , R. A. (1951). On information and sufficiency. Annals of Mathematical Statistics, 22, 79–86. Minka, T. P. (2000). Old and new matrix algebra use ...and Arabie, P. Comparing partitions. Journal of Classification, 2(1):193–218, 1985. Kullback , S. and Leibler , R. A. On information and suf- ficiency...the test input density to a lin- ear combination of class-wise input distributions under the Kullback - Leibler (KL) divergence ( Kullback
Genome-Wide Association Mapping for Intelligence in Military Working Dogs: Canine Cohort, Canine Intelligence Assessment Regimen, Genome-Wide Single Nucleotide Polymorphism (SNP) Typing, and Unsupervised Classification Algorithm for Genome-Wide Association Data Analysis

DTIC Science & Technology

2011-09-01

Almasy, L, Blangero, J. (2009) Human QTL linkage mapping. Genetica 136:333-340. Amos, CI. (2007) Successful design and conduct of genome-wide...quantitative trait loci. Genetica 136:237-243. Skol AD, Scott LJ, Abecasis GR, Boehnke M. (2006) Joint analysis is more efficient than replication
From image captioning to video summary using deep recurrent networks and unsupervised segmentation

NASA Astrophysics Data System (ADS)

Morosanu, Bogdan-Andrei; Lemnaru, Camelia

2018-04-01

Automatic captioning systems based on recurrent neural networks have been tremendously successful at providing realistic natural language captions for complex and varied image data. We explore methods for adapting existing models trained on large image caption data sets to a similar problem, that of summarising videos using natural language descriptions and frame selection. These architectures create internal high level representations of the input image that can be used to define probability distributions and distance metrics on these distributions. Specifically, we interpret each hidden unit inside a layer of the caption model as representing the un-normalised log probability of some unknown image feature of interest for the caption generation process. We can then apply well understood statistical divergence measures to express the difference between images and create an unsupervised segmentation of video frames, classifying consecutive images of low divergence as belonging to the same context, and those of high divergence as belonging to different contexts. To provide a final summary of the video, we provide a group of selected frames and a text description accompanying them, allowing a user to perform a quick exploration of large unlabeled video databases.
System for Continuous Delivery of MODIS Imagery to Internet Mapping Applications

NASA Technical Reports Server (NTRS)

Plesea, Lucian

2008-01-01

This software represents a complete, unsupervised processing chain that generates a continuously updating global image of the Earth from the most recent available MODIS Level 1B scenes. The software constantly updates a global image of the Earth at 250 m per pixel.
DOE Office of Scientific and Technical Information (OSTI.GOV)

Parekh, V; Jacobs, MA

Purpose: Multiparametric radiological imaging is used for diagnosis in patients. Potentially extracting useful features specific to a patient’s pathology would be crucial step towards personalized medicine and assessing treatment options. In order to automatically extract features directly from multiparametric radiological imaging datasets, we developed an advanced unsupervised machine learning algorithm called the multidimensional imaging radiomics-geodesics(MIRaGe). Methods: Seventy-six breast tumor patients underwent 3T MRI breast imaging were used for this study. We tested the MIRaGe algorithm to extract features for classification of breast tumors into benign or malignant. The MRI parameters used were T1-weighted, T2-weighted, dynamic contrast enhanced MR imaging (DCE-MRI)more » and diffusion weighted imaging(DWI). The MIRaGe algorithm extracted the radiomics-geodesics features (RGFs) from multiparametric MRI datasets. This enable our method to learn the intrinsic manifold representations corresponding to the patients. To determine the informative RGF, a modified Isomap algorithm(t-Isomap) was created for a radiomics-geodesics feature space(tRGFS) to avoid overfitting. Final classification was performed using SVM. The predictive power of the RGFs was tested and validated using k-fold cross validation. Results: The RGFs extracted by the MIRaGe algorithm successfully classified malignant lesions from benign lesions with a sensitivity of 93% and a specificity of 91%. The top 50 RGFs identified as the most predictive by the t-Isomap procedure were consistent with the radiological parameters known to be associated with breast cancer diagnosis and were categorized as kinetic curve characterizing RGFs, wash-in rate characterizing RGFs, wash-out rate characterizing RGFs and morphology characterizing RGFs. Conclusion: In this paper, we developed a novel feature extraction algorithm for multiparametric radiological imaging. The results demonstrated the power of the MIRaGe algorithm at automatically discovering useful feature representations directly from the raw multiparametric MRI data. In conclusion, the MIRaGe informatics model provides a powerful tool with applicability in cancer diagnosis and a possibility of extension to other kinds of pathologies. NIH (P50CA103175, 5P30CA006973 (IRAT), R01CA190299, U01CA140204), Siemens Medical Systems (JHU-2012-MR-86-01) and Nivida Graphics Corporation.« less
High Throughput Multispectral Image Processing with Applications in Food Science.

PubMed

Tsakanikas, Panagiotis; Pavlidis, Dimitris; Nychas, George-John

2015-01-01

Recently, machine vision is gaining attention in food science as well as in food industry concerning food quality assessment and monitoring. Into the framework of implementation of Process Analytical Technology (PAT) in the food industry, image processing can be used not only in estimation and even prediction of food quality but also in detection of adulteration. Towards these applications on food science, we present here a novel methodology for automated image analysis of several kinds of food products e.g. meat, vanilla crème and table olives, so as to increase objectivity, data reproducibility, low cost information extraction and faster quality assessment, without human intervention. Image processing's outcome will be propagated to the downstream analysis. The developed multispectral image processing method is based on unsupervised machine learning approach (Gaussian Mixture Models) and a novel unsupervised scheme of spectral band selection for segmentation process optimization. Through the evaluation we prove its efficiency and robustness against the currently available semi-manual software, showing that the developed method is a high throughput approach appropriate for massive data extraction from food samples.
Combining Gravitational Wave Events with their Electromagnetic Counterparts: A Realistic Joint False-Alarm Rate

NASA Astrophysics Data System (ADS)

Ackley, Kendall; Eikenberry, Stephen; Klimenko, Sergey; LIGO Team

2017-01-01

We present a false-alarm rate for a joint detection of gravitational wave (GW) events and associated electromagnetic (EM) counterparts for Advanced LIGO and Virgo (LV) observations during the first years of operation. Using simulated GW events and their recostructed probability skymaps, we tile over the error regions using sets of archival wide-field telescope survey images and recover the number of astrophysical transients to be expected during LV-EM followup. With the known GW event injection coordinates we inject artificial electromagnetic (EM) sources at that site based on theoretical and observational models on a one-to-one basis. We calculate the EM false-alarm probability using an unsupervised machine learning algorithm based on shapelet analysis which has shown to be a strong discriminator between astrophysical transients and image artifacts while reducing the set of transients to be manually vetted by five orders of magnitude. We also show the performance of our method in context with other machine-learned transient classification and reduction algorithms, showing comparability without the need for a large set of training data opening the possibility for next-generation telescopes to take advantage of this pipeline for LV-EM followup missions.
Quantification of Hepatorenal Index for Computer-Aided Fatty Liver Classification with Self-Organizing Map and Fuzzy Stretching from Ultrasonography.

PubMed

Kim, Kwang Baek; Kim, Chang Won

2015-01-01

Accurate measures of liver fat content are essential for investigating hepatic steatosis. For a noninvasive inexpensive ultrasonographic analysis, it is necessary to validate the quantitative assessment of liver fat content so that fully automated reliable computer-aided software can assist medical practitioners without any operator subjectivity. In this study, we attempt to quantify the hepatorenal index difference between the liver and the kidney with respect to the multiple severity status of hepatic steatosis. In order to do this, a series of carefully designed image processing techniques, including fuzzy stretching and edge tracking, are applied to extract regions of interest. Then, an unsupervised neural learning algorithm, the self-organizing map, is designed to establish characteristic clusters from the image, and the distribution of the hepatorenal index values with respect to the different levels of the fatty liver status is experimentally verified to estimate the differences in the distribution of the hepatorenal index. Such findings will be useful in building reliable computer-aided diagnostic software if combined with a good set of other characteristic feature sets and powerful machine learning classifiers in the future.
Quantification of Hepatorenal Index for Computer-Aided Fatty Liver Classification with Self-Organizing Map and Fuzzy Stretching from Ultrasonography

PubMed Central

Kim, Kwang Baek

2015-01-01

Accurate measures of liver fat content are essential for investigating hepatic steatosis. For a noninvasive inexpensive ultrasonographic analysis, it is necessary to validate the quantitative assessment of liver fat content so that fully automated reliable computer-aided software can assist medical practitioners without any operator subjectivity. In this study, we attempt to quantify the hepatorenal index difference between the liver and the kidney with respect to the multiple severity status of hepatic steatosis. In order to do this, a series of carefully designed image processing techniques, including fuzzy stretching and edge tracking, are applied to extract regions of interest. Then, an unsupervised neural learning algorithm, the self-organizing map, is designed to establish characteristic clusters from the image, and the distribution of the hepatorenal index values with respect to the different levels of the fatty liver status is experimentally verified to estimate the differences in the distribution of the hepatorenal index. Such findings will be useful in building reliable computer-aided diagnostic software if combined with a good set of other characteristic feature sets and powerful machine learning classifiers in the future. PMID:26247023
Characterizing Interference in Radio Astronomy Observations through Active and Unsupervised Learning

NASA Technical Reports Server (NTRS)

Doran, G.

2013-01-01

In the process of observing signals from astronomical sources, radio astronomers must mitigate the effects of manmade radio sources such as cell phones, satellites, aircraft, and observatory equipment. Radio frequency interference (RFI) often occurs as short bursts (< 1 ms) across a broad range of frequencies, and can be confused with signals from sources of interest such as pulsars. With ever-increasing volumes of data being produced by observatories, automated strategies are required to detect, classify, and characterize these short "transient" RFI events. We investigate an active learning approach in which an astronomer labels events that are most confusing to a classifier, minimizing the human effort required for classification. We also explore the use of unsupervised clustering techniques, which automatically group events into classes without user input. We apply these techniques to data from the Parkes Multibeam Pulsar Survey to characterize several million detected RFI events from over a thousand hours of observation.
CHISSL: A Human-Machine Collaboration Space for Unsupervised Learning

DOE Office of Scientific and Technical Information (OSTI.GOV)

Arendt, Dustin L.; Komurlu, Caner; Blaha, Leslie M.

We developed CHISSL, a human-machine interface that utilizes supervised machine learning in an unsupervised context to help the user group unlabeled instances by her own mental model. The user primarily interacts via correction (moving a misplaced instance into its correct group) or confirmation (accepting that an instance is placed in its correct group). Concurrent with the user's interactions, CHISSL trains a classification model guided by the user's grouping of the data. It then predicts the group of unlabeled instances and arranges some of these alongside the instances manually organized by the user. We hypothesize that this mode of human andmore » machine collaboration is more effective than Active Learning, wherein the machine decides for itself which instances should be labeled by the user. We found supporting evidence for this hypothesis in a pilot study where we applied CHISSL to organize a collection of handwritten digits.« less
Unsupervised segmentation with dynamical units.

PubMed

Rao, A Ravishankar; Cecchi, Guillermo A; Peck, Charles C; Kozloski, James R

2008-01-01

In this paper, we present a novel network to separate mixtures of inputs that have been previously learned. A significant capability of the network is that it segments the components of each input object that most contribute to its classification. The network consists of amplitude-phase units that can synchronize their dynamics, so that separation is determined by the amplitude of units in an output layer, and segmentation by phase similarity between input and output layer units. Learning is unsupervised and based on a Hebbian update, and the architecture is very simple. Moreover, efficient segmentation can be achieved even when there is considerable superposition of the inputs. The network dynamics are derived from an objective function that rewards sparse coding in the generalized amplitude-phase variables. We argue that this objective function can provide a possible formal interpretation of the binding problem and that the implementation of the network architecture and dynamics is biologically plausible.
A Gaussian mixture model based adaptive classifier for fNIRS brain-computer interfaces and its testing via simulation

NASA Astrophysics Data System (ADS)

Li, Zheng; Jiang, Yi-han; Duan, Lian; Zhu, Chao-zhe

2017-08-01

Objective. Functional near infra-red spectroscopy (fNIRS) is a promising brain imaging technology for brain-computer interfaces (BCI). Future clinical uses of fNIRS will likely require operation over long time spans, during which neural activation patterns may change. However, current decoders for fNIRS signals are not designed to handle changing activation patterns. The objective of this study is to test via simulations a new adaptive decoder for fNIRS signals, the Gaussian mixture model adaptive classifier (GMMAC). Approach. GMMAC can simultaneously classify and track activation pattern changes without the need for ground-truth labels. This adaptive classifier uses computationally efficient variational Bayesian inference to label new data points and update mixture model parameters, using the previous model parameters as priors. We test GMMAC in simulations in which neural activation patterns change over time and compare to static decoders and unsupervised adaptive linear discriminant analysis classifiers. Main results. Our simulation experiments show GMMAC can accurately decode under time-varying activation patterns: shifts of activation region, expansions of activation region, and combined contractions and shifts of activation region. Furthermore, the experiments show the proposed method can track the changing shape of the activation region. Compared to prior work, GMMAC performed significantly better than the other unsupervised adaptive classifiers on a difficult activation pattern change simulation: 99% versus <54% in two-choice classification accuracy. Significance. We believe GMMAC will be useful for clinical fNIRS-based brain-computer interfaces, including neurofeedback training systems, where operation over long time spans is required.
Deep learning-based subdivision approach for large scale macromolecules structure recovery from electron cryo tomograms

PubMed Central

Xu, Min; Chai, Xiaoqi; Muthakana, Hariank; Liang, Xiaodan; Yang, Ge; Zeev-Ben-Mordehai, Tzviya; Xing, Eric P.

2017-01-01

Abstract Motivation: Cellular Electron CryoTomography (CECT) enables 3D visualization of cellular organization at near-native state and in sub-molecular resolution, making it a powerful tool for analyzing structures of macromolecular complexes and their spatial organizations inside single cells. However, high degree of structural complexity together with practical imaging limitations makes the systematic de novo discovery of structures within cells challenging. It would likely require averaging and classifying millions of subtomograms potentially containing hundreds of highly heterogeneous structural classes. Although it is no longer difficult to acquire CECT data containing such amount of subtomograms due to advances in data acquisition automation, existing computational approaches have very limited scalability or discrimination ability, making them incapable of processing such amount of data. Results: To complement existing approaches, in this article we propose a new approach for subdividing subtomograms into smaller but relatively homogeneous subsets. The structures in these subsets can then be separately recovered using existing computation intensive methods. Our approach is based on supervised structural feature extraction using deep learning, in combination with unsupervised clustering and reference-free classification. Our experiments show that, compared with existing unsupervised rotation invariant feature and pose-normalization based approaches, our new approach achieves significant improvements in both discrimination ability and scalability. More importantly, our new approach is able to discover new structural classes and recover structures that do not exist in training data. Availability and Implementation: Source code freely available at http://www.cs.cmu.edu/∼mxu1/software. Contact: mxu1@cs.cmu.edu Supplementary information: Supplementary data are available at Bioinformatics online. PMID:28881965

Satellite altimetry in sea ice regions - detecting open water for estimating sea surface heights

NASA Astrophysics Data System (ADS)

Müller, Felix L.; Dettmering, Denise; Bosch, Wolfgang

2017-04-01

The Greenland Sea and the Farm Strait are transporting sea ice from the central Arctic ocean southwards. They are covered by a dynamic changing sea ice layer with significant influences on the Earth climate system. Between the sea ice there exist various sized open water areas known as leads, straight lined open water areas, and polynyas exhibiting a circular shape. Identifying these leads by satellite altimetry enables the extraction of sea surface height information. Analyzing the radar echoes, also called waveforms, provides information on the surface backscatter characteristics. For example waveforms reflected by calm water have a very narrow and single-peaked shape. Waveforms reflected by sea ice show more variability due to diffuse scattering. Here we analyze altimeter waveforms from different conventional pulse-limited satellite altimeters to separate open water and sea ice waveforms. An unsupervised classification approach employing partitional clustering algorithms such as K-medoids and memory-based classification methods such as K-nearest neighbor is used. The classification is based on six parameters derived from the waveform's shape, for example the maximum power or the peak's width. The open-water detection is quantitatively compared to SAR images processed while accounting for sea ice motion. The classification results are used to derive information about the temporal evolution of sea ice extent and sea surface heights. They allow to provide evidence on climate change relevant influences as for example Arctic sea level rise due to enhanced melting rates of Greenland's glaciers and an increasing fresh water influx into the Arctic ocean. Additionally, the sea ice cover extent analyzed over a long-time period provides an important indicator for a globally changing climate system.
Quasi-Supervised Scoring of Human Sleep in Polysomnograms Using Augmented Input Variables

PubMed Central

Yaghouby, Farid; Sunderam, Sridhar

2015-01-01

The limitations of manual sleep scoring make computerized methods highly desirable. Scoring errors can arise from human rater uncertainty or inter-rater variability. Sleep scoring algorithms either come as supervised classifiers that need scored samples of each state to be trained, or as unsupervised classifiers that use heuristics or structural clues in unscored data to define states. We propose a quasi-supervised classifier that models observations in an unsupervised manner but mimics a human rater wherever training scores are available. EEG, EMG, and EOG features were extracted in 30s epochs from human-scored polysomnograms recorded from 42 healthy human subjects (18 to 79 years) and archived in an anonymized, publicly accessible database. Hypnograms were modified so that: 1. Some states are scored but not others; 2. Samples of all states are scored but not for transitional epochs; and 3. Two raters with 67% agreement are simulated. A framework for quasi-supervised classification was devised in which unsupervised statistical models—specifically Gaussian mixtures and hidden Markov models—are estimated from unlabeled training data, but the training samples are augmented with variables whose values depend on available scores. Classifiers were fitted to signal features incorporating partial scores, and used to predict scores for complete recordings. Performance was assessed using Cohen's K statistic. The quasi-supervised classifier performed significantly better than an unsupervised model and sometimes as well as a completely supervised model despite receiving only partial scores. The quasi-supervised algorithm addresses the need for classifiers that mimic scoring patterns of human raters while compensating for their limitations. PMID:25679475
Quasi-supervised scoring of human sleep in polysomnograms using augmented input variables.

PubMed

Yaghouby, Farid; Sunderam, Sridhar

2015-04-01

The limitations of manual sleep scoring make computerized methods highly desirable. Scoring errors can arise from human rater uncertainty or inter-rater variability. Sleep scoring algorithms either come as supervised classifiers that need scored samples of each state to be trained, or as unsupervised classifiers that use heuristics or structural clues in unscored data to define states. We propose a quasi-supervised classifier that models observations in an unsupervised manner but mimics a human rater wherever training scores are available. EEG, EMG, and EOG features were extracted in 30s epochs from human-scored polysomnograms recorded from 42 healthy human subjects (18-79 years) and archived in an anonymized, publicly accessible database. Hypnograms were modified so that: 1. Some states are scored but not others; 2. Samples of all states are scored but not for transitional epochs; and 3. Two raters with 67% agreement are simulated. A framework for quasi-supervised classification was devised in which unsupervised statistical models-specifically Gaussian mixtures and hidden Markov models--are estimated from unlabeled training data, but the training samples are augmented with variables whose values depend on available scores. Classifiers were fitted to signal features incorporating partial scores, and used to predict scores for complete recordings. Performance was assessed using Cohen's Κ statistic. The quasi-supervised classifier performed significantly better than an unsupervised model and sometimes as well as a completely supervised model despite receiving only partial scores. The quasi-supervised algorithm addresses the need for classifiers that mimic scoring patterns of human raters while compensating for their limitations. Copyright © 2015 Elsevier Ltd. All rights reserved.
The Danish Fracture Database can monitor quality of fracture-related surgery, surgeons' experience level and extent of supervision.

PubMed

Andersen, Morten Jon; Gromov, Kiril; Brix, Michael; Troelsen, Anders

2014-06-01

The importance of supervision and of surgeons' level of experience in relation to patient outcome have been demonstrated in both hip fracture and arthroplasty surgery. The aim of this study was to describe the surgeons' experience level and the extent of supervision for: 1) fracture-related surgery in general; 2) the three most frequent primary operations and reoperations; and 3) primary operations during and outside regular working hours. A total of 9,767 surgical procedures were identified from the Danish Fracture Database (DFDB). Procedures were grouped based on the surgeons' level of experience, extent of supervision, type (primary, planned secondary or reoperation), classification (AO Müller), and whether they were performed during or outside regular hours. Interns and junior residents combined performed 46% of all procedures. A total of 90% of surgeries by interns were performed under supervision, whereas 32% of operations by junior residents were unsupervised. Supervision was absent in 14-16% and 22-33% of the three most frequent primary procedures and reoperations when performed by interns and junior residents, respectively. The proportion of unsupervised procedures by junior residents grew from 30% during to 40% (p < 0.001) outside regular hours. Interns and junior residents together performed almost half of all fracture-related surgery. The extent of supervision was generally high; however, a third of the primary procedures performed by junior residents were unsupervised. The extent of unsupervised surgery performed by junior residents was significantly higher outside regular hours. not relevant. The Danish Fracture Database ("Dansk Frakturdatabase") was approved by the Danish Data Protection Agency ID: 01321.
Multitemporal Snow Cover Mapping in Mountainous Terrain for Landsat Climate Data Record Development

NASA Technical Reports Server (NTRS)

Crawford, Christopher J.; Manson, Steven M.; Bauer, Marvin E.; Hall, Dorothy K.

2013-01-01

A multitemporal method to map snow cover in mountainous terrain is proposed to guide Landsat climate data record (CDR) development. The Landsat image archive including MSS, TM, and ETM+ imagery was used to construct a prototype Landsat snow cover CDR for the interior northwestern United States. Landsat snow cover CDRs are designed to capture snow-covered area (SCA) variability at discrete bi-monthly intervals that correspond to ground-based snow telemetry (SNOTEL) snow-water-equivalent (SWE) measurements. The June 1 bi-monthly interval was selected for initial CDR development, and was based on peak snowmelt timing for this mountainous region. Fifty-four Landsat images from 1975 to 2011 were preprocessed that included image registration, top-of-the-atmosphere (TOA) reflectance conversion, cloud and shadow masking, and topographic normalization. Snow covered pixels were retrieved using the normalized difference snow index (NDSI) and unsupervised classification, and pixels having greater (less) than 50% snow cover were classified presence (absence). A normalized SCA equation was derived to independently estimate SCA given missing image coverage and cloud-shadow contamination. Relative frequency maps of missing pixels were assembled to assess whether systematic biases were embedded within this Landsat CDR. Our results suggest that it is possible to confidently estimate historical bi-monthly SCA from partially cloudy Landsat images. This multitemporal method is intended to guide Landsat CDR development for freshwaterscarce regions of the western US to monitor climate-driven changes in mountain snowpack extent.
Gene expression-based molecular diagnostic system for malignant gliomas is superior to histological diagnosis.

PubMed

Shirahata, Mitsuaki; Iwao-Koizumi, Kyoko; Saito, Sakae; Ueno, Noriko; Oda, Masashi; Hashimoto, Nobuo; Takahashi, Jun A; Kato, Kikuya

2007-12-15

Current morphology-based glioma classification methods do not adequately reflect the complex biology of gliomas, thus limiting their prognostic ability. In this study, we focused on anaplastic oligodendroglioma and glioblastoma, which typically follow distinct clinical courses. Our goal was to construct a clinically useful molecular diagnostic system based on gene expression profiling. The expression of 3,456 genes in 32 patients, 12 and 20 of whom had prognostically distinct anaplastic oligodendroglioma and glioblastoma, respectively, was measured by PCR array. Next to unsupervised methods, we did supervised analysis using a weighted voting algorithm to construct a diagnostic system discriminating anaplastic oligodendroglioma from glioblastoma. The diagnostic accuracy of this system was evaluated by leave-one-out cross-validation. The clinical utility was tested on a microarray-based data set of 50 malignant gliomas from a previous study. Unsupervised analysis showed divergent global gene expression patterns between the two tumor classes. A supervised binary classification model showed 100% (95% confidence interval, 89.4-100%) diagnostic accuracy by leave-one-out cross-validation using 168 diagnostic genes. Applied to a gene expression data set from a previous study, our model correlated better with outcome than histologic diagnosis, and also displayed 96.6% (28 of 29) consistency with the molecular classification scheme used for these histologically controversial gliomas in the original article. Furthermore, we observed that histologically diagnosed glioblastoma samples that shared anaplastic oligodendroglioma molecular characteristics tended to be associated with longer survival. Our molecular diagnostic system showed reproducible clinical utility and prognostic ability superior to traditional histopathologic diagnosis for malignant glioma.
Identification of chronic rhinosinusitis phenotypes using cluster analysis.

PubMed

Soler, Zachary M; Hyer, J Madison; Ramakrishnan, Viswanathan; Smith, Timothy L; Mace, Jess; Rudmik, Luke; Schlosser, Rodney J

2015-05-01

Current clinical classifications of chronic rhinosinusitis (CRS) have been largely defined based upon preconceived notions of factors thought to be important, such as polyp or eosinophil status. Unfortunately, these classification systems have little correlation with symptom severity or treatment outcomes. Unsupervised clustering can be used to identify phenotypic subgroups of CRS patients, describe clinical differences in these clusters and define simple algorithms for classification. A multi-institutional, prospective study of 382 patients with CRS who had failed initial medical therapy completed the Sino-Nasal Outcome Test (SNOT-22), Rhinosinusitis Disability Index (RSDI), Medical Outcomes Study Short Form-12 (SF-12), Pittsburgh Sleep Quality Index (PSQI), and Patient Health Questionnaire (PHQ-2). Objective measures of CRS severity included Brief Smell Identification Test (B-SIT), CT, and endoscopy scoring. All variables were reduced and unsupervised hierarchical clustering was performed. After clusters were defined, variations in medication usage were analyzed. Discriminant analysis was performed to develop a simplified, clinically useful algorithm for clustering. Clustering was largely determined by age, severity of patient reported outcome measures, depression, and fibromyalgia. CT and endoscopy varied somewhat among clusters. Traditional clinical measures, including polyp/atopic status, prior surgery, B-SIT and asthma, did not vary among clusters. A simplified algorithm based upon productivity loss, SNOT-22 score, and age predicted clustering with 89% accuracy. Medication usage among clusters did vary significantly. A simplified algorithm based upon hierarchical clustering is able to classify CRS patients and predict medication usage. Further studies are warranted to determine if such clustering predicts treatment outcomes. © 2015 ARS-AAOA, LLC.
Heterogeneous activation of the TGFβ pathway in glioblastomas identified by gene expression-based classification using TGFβ-responsive genes

PubMed Central

Xu, Xie L; Kapoun, Ann M

2009-01-01

Background TGFβ has emerged as an attractive target for the therapeutic intervention of glioblastomas. Aberrant TGFβ overproduction in glioblastoma and other high-grade gliomas has been reported, however, to date, none of these reports has systematically examined the components of TGFβ signaling to gain a comprehensive view of TGFβ activation in large cohorts of human glioma patients. Methods TGFβ activation in mammalian cells leads to a transcriptional program that typically affects 5–10% of the genes in the genome. To systematically examine the status of TGFβ activation in high-grade glial tumors, we compiled a gene set of transcriptional response to TGFβ stimulation from tissue culture and in vivo animal studies. These genes were used to examine the status of TGFβ activation in high-grade gliomas including a large cohort of glioblastomas. Unsupervised and supervised classification analysis was performed in two independent, publicly available glioma microarray datasets. Results Unsupervised and supervised classification using the TGFβ-responsive gene list in two independent glial tumor gene expression data sets revealed various levels of TGFβ activation in these tumors. Among glioblastomas, one of the most devastating human cancers, two subgroups were identified that showed distinct TGFβ activation patterns as measured from transcriptional responses. Approximately 62% of glioblastoma samples analyzed showed strong TGFβ activation, while the rest showed a weak TGFβ transcriptional response. Conclusion Our findings suggest heterogeneous TGFβ activation in glioblastomas, which may cause potential differences in responses to anti-TGFβ therapies in these two distinct subgroups of glioblastomas patients. PMID:19192267
LICRE: unsupervised feature correlation reduction for lipidomics.

PubMed

Wong, Gerard; Chan, Jeffrey; Kingwell, Bronwyn A; Leckie, Christopher; Meikle, Peter J

2014-10-01

Recent advances in high-throughput lipid profiling by liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) have made it possible to quantify hundreds of individual molecular lipid species (e.g. fatty acyls, glycerolipids, glycerophospholipids, sphingolipids) in a single experimental run for hundreds of samples. This enables the lipidome of large cohorts of subjects to be profiled to identify lipid biomarkers significantly associated with disease risk, progression and treatment response. Clinically, these lipid biomarkers can be used to construct classification models for the purpose of disease screening or diagnosis. However, the inclusion of a large number of highly correlated biomarkers within a model may reduce classification performance, unnecessarily inflate associated costs of a diagnosis or a screen and reduce the feasibility of clinical translation. An unsupervised feature reduction approach can reduce feature redundancy in lipidomic biomarkers by limiting the number of highly correlated lipids while retaining informative features to achieve good classification performance for various clinical outcomes. Good predictive models based on a reduced number of biomarkers are also more cost effective and feasible from a clinical translation perspective. The application of LICRE to various lipidomic datasets in diabetes and cardiovascular disease demonstrated superior discrimination in terms of the area under the receiver operator characteristic curve while using fewer lipid markers when predicting various clinical outcomes. The MATLAB implementation of LICRE is available from http://ww2.cs.mu.oz.au/∼gwong/LICRE © The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.
Rough-Fuzzy Clustering and Unsupervised Feature Selection for Wavelet Based MR Image Segmentation

PubMed Central

Maji, Pradipta; Roy, Shaswati

2015-01-01

Image segmentation is an indispensable process in the visualization of human tissues, particularly during clinical analysis of brain magnetic resonance (MR) images. For many human experts, manual segmentation is a difficult and time consuming task, which makes an automated brain MR image segmentation method desirable. In this regard, this paper presents a new segmentation method for brain MR images, integrating judiciously the merits of rough-fuzzy computing and multiresolution image analysis technique. The proposed method assumes that the major brain tissues, namely, gray matter, white matter, and cerebrospinal fluid from the MR images are considered to have different textural properties. The dyadic wavelet analysis is used to extract the scale-space feature vector for each pixel, while the rough-fuzzy clustering is used to address the uncertainty problem of brain MR image segmentation. An unsupervised feature selection method is introduced, based on maximum relevance-maximum significance criterion, to select relevant and significant textural features for segmentation problem, while the mathematical morphology based skull stripping preprocessing step is proposed to remove the non-cerebral tissues like skull. The performance of the proposed method, along with a comparison with related approaches, is demonstrated on a set of synthetic and real brain MR images using standard validity indices. PMID:25848961
Unsupervised detection and removal of muscle artifacts from scalp EEG recordings using canonical correlation analysis, wavelets and random forests.

PubMed

Anastasiadou, Maria N; Christodoulakis, Manolis; Papathanasiou, Eleftherios S; Papacostas, Savvas S; Mitsis, Georgios D

2017-09-01

This paper proposes supervised and unsupervised algorithms for automatic muscle artifact detection and removal from long-term EEG recordings, which combine canonical correlation analysis (CCA) and wavelets with random forests (RF). The proposed algorithms first perform CCA and continuous wavelet transform of the canonical components to generate a number of features which include component autocorrelation values and wavelet coefficient magnitude values. A subset of the most important features is subsequently selected using RF and labelled observations (supervised case) or synthetic data constructed from the original observations (unsupervised case). The proposed algorithms are evaluated using realistic simulation data as well as 30min epochs of non-invasive EEG recordings obtained from ten patients with epilepsy. We assessed the performance of the proposed algorithms using classification performance and goodness-of-fit values for noisy and noise-free signal windows. In the simulation study, where the ground truth was known, the proposed algorithms yielded almost perfect performance. In the case of experimental data, where expert marking was performed, the results suggest that both the supervised and unsupervised algorithm versions were able to remove artifacts without affecting noise-free channels considerably, outperforming standard CCA, independent component analysis (ICA) and Lagged Auto-Mutual Information Clustering (LAMIC). The proposed algorithms achieved excellent performance for both simulation and experimental data. Importantly, for the first time to our knowledge, we were able to perform entirely unsupervised artifact removal, i.e. without using already marked noisy data segments, achieving performance that is comparable to the supervised case. Overall, the results suggest that the proposed algorithms yield significant future potential for improving EEG signal quality in research or clinical settings without the need for marking by expert neurophysiologists, EMG signal recording and user visual inspection. Copyright © 2017 International Federation of Clinical Neurophysiology. Published by Elsevier B.V. All rights reserved.
Integrated Use of Remote Sensing, GIS and GPS Technology for Monitoring the Environmental Problem of Shyamnagar

NASA Astrophysics Data System (ADS)

Akbar, M. S.; Sarker, M. H.; Sattar, M. A.; Sarwar, G. M.; Rahman, S. M. M.; Rahman, M. M.; Khan, Z. U.

2017-05-01

Cultivation of shrimp mostly in unplanned way has been considered as one of the major environmental disasters of Shamnagar. Villagers surrounding the rivers are mainly involved with fish (shrimp) cultivation. So, fertile agriculture land has been converted to shrimp cultivation. Conversion of agriculture land to other usage is a common but acute problem for land resources of the country like Bangladesh. Conventional methods for collecting this information are relatively costly and time consuming. Contrarily, Remote Sensing satellite observation with its unique capability to provide cost-effective support in compiling the latest information about the natural resource. Remote sensing, in conjunction with GIS, has been widely applied and been recognized as a powerful and effective tool in detecting land use and land cover changes. RapidEye, Landsat8 images were used to identify land use and land cover of the area during the period 2008 and 2015. Google images were used to identify the micro-level land use features of the same period. Multi-spectral classifications using unsupervised and supervised classification were done and results have been compared based on the field investigation. The study reveals that during the period 2008 to 2015 agricultural practice has been reduced from 35 % to 21 % and shrimp cultivation area increased from 38 % to 50 %. Due to the impact of high salinity and salt water intrusion caused by natural disaster, agricultural activities is reduced and farmers have been converted to other practices, as a result shrimp farming is gaining popularity in the area.
A new paradigm of oral cancer detection using digital infrared thermal imaging

NASA Astrophysics Data System (ADS)

Chakraborty, M.; Mukhopadhyay, S.; Dasgupta, A.; Banerjee, S.; Mukhopadhyay, S.; Patsa, S.; Ray, J. G.; Chaudhuri, K.

2016-03-01

Histopathology is considered the gold standard for oral cancer detection. But a major fraction of patient pop- ulation is incapable of accessing such healthcare facilities due to poverty. Moreover, such analysis may report false negatives when test tissue is not collected from exact cancerous location. The proposed work introduces a pioneering computer aided paradigm of fast, non-invasive and non-ionizing modality for oral cancer detection us- ing Digital Infrared Thermal Imaging (DITI). Due to aberrant metabolic activities in carcinogenic facial regions, heat signatures of patients are different from that of normal subjects. The proposed work utilizes asymmetry of temperature distribution of facial regions as principle cue for cancer detection. Three views of a subject, viz. front, left and right are acquired using long infrared (7:5 - 13μm) camera for analysing distribution of temperature. We study asymmetry of facial temperature distribution between: a) left and right profile faces and b) left and right half of frontal face. Comparison of temperature distribution suggests that patients manifest greater asymmetry compared to normal subjects. For classification, we initially use k-means and fuzzy k-means for unsupervised clustering followed by cluster class prototype assignment based on majority voting. Average classification accuracy of 91:5% and 92:8% are achieved by k-mean and fuzzy k-mean framework for frontal face. The corresponding metrics for profile face are 93:4% and 95%. Combining features of frontal and profile faces, average accuracies are increased to 96:2% and 97:6% respectively for k-means and fuzzy k-means framework.
Applicability Assessment of Uavsar Data in Wetland Monitoring: a Case Study of Louisiana Wetland

NASA Astrophysics Data System (ADS)

Zhao, J.; Niu, Y.; Lu, Z.; Yang, J.; Li, P.; Liu, W.

2018-04-01

Wetlands are highly productive and support a wide variety of ecosystem goods and services. Monitoring wetland is essential and potential. Because of the repeat-pass nature of satellite orbit and airborne, time-series of remote sensing data can be obtained to monitor wetland. UAVSAR is a NASA L-band synthetic aperture radar (SAR) sensor compact pod-mounted polarimetric instrument for interferometric repeat-track observations. Moreover, UAVSAR images can accurately map crustal deformations associated with natural hazards, such as volcanoes and earthquakes. And its polarization agility facilitates terrain and land-use classification and change detection. In this paper, the multi-temporal UAVSAR data are applied for monitoring the wetland change. Using the multi-temporal polarimetric SAR (PolSAR) data, the change detection maps are obtained by unsupervised and supervised method. And the coherence is extracted from the interfometric SAR (InSAR) data to verify the accuracy of change detection map. The experimental results show that the multi-temporal UAVSAR data is fit for wetland monitor.
Using Wikipedia to learn semantic feature representations of concrete concepts in neuroimaging experiments

PubMed Central

Pereira, Francisco; Botvinick, Matthew; Detre, Greg

2012-01-01

In this paper we show that a corpus of a few thousand Wikipedia articles about concrete or visualizable concepts can be used to produce a low-dimensional semantic feature representation of those concepts. The purpose of such a representation is to serve as a model of the mental context of a subject during functional magnetic resonance imaging (fMRI) experiments. A recent study [19] showed that it was possible to predict fMRI data acquired while subjects thought about a concrete concept, given a representation of those concepts in terms of semantic features obtained with human supervision. We use topic models on our corpus to learn semantic features from text in an unsupervised manner, and show that those features can outperform those in [19] in demanding 12-way and 60-way classification tasks. We also show that these features can be used to uncover similarity relations in brain activation for different concepts which parallel those relations in behavioral data from human subjects. PMID:23243317
Clustering performance comparison using K-means and expectation maximization algorithms.

PubMed

Jung, Yong Gyu; Kang, Min Soo; Heo, Jun

2014-11-14

Clustering is an important means of data mining based on separating data categories by similar features. Unlike the classification algorithm, clustering belongs to the unsupervised type of algorithms. Two representatives of the clustering algorithms are the K -means and the expectation maximization (EM) algorithm. Linear regression analysis was extended to the category-type dependent variable, while logistic regression was achieved using a linear combination of independent variables. To predict the possibility of occurrence of an event, a statistical approach is used. However, the classification of all data by means of logistic regression analysis cannot guarantee the accuracy of the results. In this paper, the logistic regression analysis is applied to EM clusters and the K -means clustering method for quality assessment of red wine, and a method is proposed for ensuring the accuracy of the classification results.
Eigenvalue-eigenvector decomposition (EED) analysis of dissimilarity and covariance matrix obtained from total synchronous fluorescence spectral (TSFS) data sets of herbal preparations: Optimizing the classification approach.

PubMed

Tarai, Madhumita; Kumar, Keshav; Divya, O; Bairi, Partha; Mishra, Kishor Kumar; Mishra, Ashok Kumar

2017-09-05

The present work compares the dissimilarity and covariance based unsupervised chemometric classification approaches by taking the total synchronous fluorescence spectroscopy data sets acquired for the cumin and non-cumin based herbal preparations. The conventional decomposition method involves eigenvalue-eigenvector analysis of the covariance of the data set and finds the factors that can explain the overall major sources of variation present in the data set. The conventional approach does this irrespective of the fact that the samples belong to intrinsically different groups and hence leads to poor class separation. The present work shows that classification of such samples can be optimized by performing the eigenvalue-eigenvector decomposition on the pair-wise dissimilarity matrix. Copyright © 2017 Elsevier B.V. All rights reserved.
Unsupervised change detection of multispectral images based on spatial constraint chi-squared transform and Markov random field model

NASA Astrophysics Data System (ADS)

Shi, Aiye; Wang, Chao; Shen, Shaohong; Huang, Fengchen; Ma, Zhenli

2016-10-01

Chi-squared transform (CST), as a statistical method, can describe the difference degree between vectors. The CST-based methods operate directly on information stored in the difference image and are simple and effective methods for detecting changes in remotely sensed images that have been registered and aligned. However, the technique does not take spatial information into consideration, which leads to much noise in the result of change detection. An improved unsupervised change detection method is proposed based on spatial constraint CST (SCCST) in combination with a Markov random field (MRF) model. First, the mean and variance matrix of the difference image of bitemporal images are estimated by an iterative trimming method. In each iteration, spatial information is injected to reduce scattered changed points (also known as "salt and pepper" noise). To determine the key parameter confidence level in the SCCST method, a pseudotraining dataset is constructed to estimate the optimal value. Then, the result of SCCST, as an initial solution of change detection, is further improved by the MRF model. The experiments on simulated and real multitemporal and multispectral images indicate that the proposed method performs well in comprehensive indices compared with other methods.
Stratification Learning: Detecting Mixed Density and Dimensionality in High Dimensional Point Clouds (PREPRINT)

DTIC Science & Technology

2006-09-01

Medioni, [11], estimates the local dimension using tensor voting . These recent works have clearly shown the necessity to go beyond manifold learning, into...2005. [11] P. Mordohai and G. Medioni. Unsupervised dimensionality estimation and manifold learning in high-dimensional spaces by tensor voting . In...walking, jumping, and arms waving. The whole run took 361 seconds in Matlab , while the classification time (PMM) can be neglected compared to the kNN
Unsupervised background-constrained tank segmentation of infrared images in complex background based on the Otsu method.

PubMed

Zhou, Yulong; Gao, Min; Fang, Dan; Zhang, Baoquan

2016-01-01

In an effort to implement fast and effective tank segmentation from infrared images in complex background, the threshold of the maximum between-class variance method (i.e., the Otsu method) is analyzed and the working mechanism of the Otsu method is discussed. Subsequently, a fast and effective method for tank segmentation from infrared images in complex background is proposed based on the Otsu method via constraining the complex background of the image. Considering the complexity of background, the original image is firstly divided into three classes of target region, middle background and lower background via maximizing the sum of their between-class variances. Then, the unsupervised background constraint is implemented based on the within-class variance of target region and hence the original image can be simplified. Finally, the Otsu method is applied to simplified image for threshold selection. Experimental results on a variety of tank infrared images (880 × 480 pixels) in complex background demonstrate that the proposed method enjoys better segmentation performance and even could be comparative with the manual segmentation in segmented results. In addition, its average running time is only 9.22 ms, implying the new method with good performance in real time processing.

A comparison of performance of automatic cloud coverage assessment algorithm for Formosat-2 image using clustering-based and spatial thresholding methods

NASA Astrophysics Data System (ADS)

Hsu, Kuo-Hsien

2012-11-01

Formosat-2 image is a kind of high-spatial-resolution (2 meters GSD) remote sensing satellite data, which includes one panchromatic band and four multispectral bands (Blue, Green, Red, near-infrared). An essential sector in the daily processing of received Formosat-2 image is to estimate the cloud statistic of image using Automatic Cloud Coverage Assessment (ACCA) algorithm. The information of cloud statistic of image is subsequently recorded as an important metadata for image product catalog. In this paper, we propose an ACCA method with two consecutive stages: preprocessing and post-processing analysis. For pre-processing analysis, the un-supervised K-means classification, Sobel's method, thresholding method, non-cloudy pixels reexamination, and cross-band filter method are implemented in sequence for cloud statistic determination. For post-processing analysis, Box-Counting fractal method is implemented. In other words, the cloud statistic is firstly determined via pre-processing analysis, the correctness of cloud statistic of image of different spectral band is eventually cross-examined qualitatively and quantitatively via post-processing analysis. The selection of an appropriate thresholding method is very critical to the result of ACCA method. Therefore, in this work, We firstly conduct a series of experiments of the clustering-based and spatial thresholding methods that include Otsu's, Local Entropy(LE), Joint Entropy(JE), Global Entropy(GE), and Global Relative Entropy(GRE) method, for performance comparison. The result shows that Otsu's and GE methods both perform better than others for Formosat-2 image. Additionally, our proposed ACCA method by selecting Otsu's method as the threshoding method has successfully extracted the cloudy pixels of Formosat-2 image for accurate cloud statistic estimation.
Spectral embedding finds meaningful (relevant) structure in image and microarray data

PubMed Central

Higgs, Brandon W; Weller, Jennifer; Solka, Jeffrey L

2006-01-01

Background Accurate methods for extraction of meaningful patterns in high dimensional data have become increasingly important with the recent generation of data types containing measurements across thousands of variables. Principal components analysis (PCA) is a linear dimensionality reduction (DR) method that is unsupervised in that it relies only on the data; projections are calculated in Euclidean or a similar linear space and do not use tuning parameters for optimizing the fit to the data. However, relationships within sets of nonlinear data types, such as biological networks or images, are frequently mis-rendered into a low dimensional space by linear methods. Nonlinear methods, in contrast, attempt to model important aspects of the underlying data structure, often requiring parameter(s) fitting to the data type of interest. In many cases, the optimal parameter values vary when different classification algorithms are applied on the same rendered subspace, making the results of such methods highly dependent upon the type of classifier implemented. Results We present the results of applying the spectral method of Lafon, a nonlinear DR method based on the weighted graph Laplacian, that minimizes the requirements for such parameter optimization for two biological data types. We demonstrate that it is successful in determining implicit ordering of brain slice image data and in classifying separate species in microarray data, as compared to two conventional linear methods and three nonlinear methods (one of which is an alternative spectral method). This spectral implementation is shown to provide more meaningful information, by preserving important relationships, than the methods of DR presented for comparison. Tuning parameter fitting is simple and is a general, rather than data type or experiment specific approach, for the two datasets analyzed here. Tuning parameter optimization is minimized in the DR step to each subsequent classification method, enabling the possibility of valid cross-experiment comparisons. Conclusion Results from the spectral method presented here exhibit the desirable properties of preserving meaningful nonlinear relationships in lower dimensional space and requiring minimal parameter fitting, providing a useful algorithm for purposes of visualization and classification across diverse datasets, a common challenge in systems biology. PMID:16483359
An object-based image analysis approach for aquaculture ponds precise mapping and monitoring: a case study of Tam Giang-Cau Hai Lagoon, Vietnam.

PubMed

Virdis, Salvatore Gonario Pasquale

2014-01-01

Monitoring and mapping shrimp farms, including their impact on land cover and land use, is critical to the sustainable management and planning of coastal zones. In this work, a methodology was proposed to set up a cost-effective and reproducible procedure that made use of satellite remote sensing, object-based classification approach, and open-source software for mapping aquaculture areas with high planimetric and thematic accuracy between 2005 and 2008. The analysis focused on two characteristic areas of interest of the Tam Giang-Cau Hai Lagoon (in central Vietnam), which have similar farming systems to other coastal aquaculture worldwide: the first was primarily characterised by locally referred "low tide" shrimp ponds, which are partially submerged areas; the second by earthed shrimp ponds, locally referred to as "high tide" ponds, which are non-submerged areas on the lagoon coast. The approach was based on the region-growing segmentation of high- and very high-resolution panchromatic images, SPOT5 and Worldview-1, and the unsupervised clustering classifier ISOSEG embedded on SPRING non-commercial software. The results, the accuracy of which was tested with a field-based aquaculture inventory, showed that in favourable situations (high tide shrimp ponds), the classification results provided high rates of accuracy (>95 %) through a fully automatic object-based classification. In unfavourable situations (low tide shrimp ponds), the performance degraded due to the low contrast between the water and the pond embankments. In these situations, the automatic results were improved by manual delineation of the embankments. Worldview-1 necessarily showed better thematic accuracy, and precise maps have been realised at a scale of up to 1:2,000. However, SPOT5 provided comparable results in terms of number of correctly classified ponds, but less accurate results in terms of the precision of mapped features. The procedure also demonstrated high degrees of reproducibility because it was applied to images with different spatial resolutions in an area that, during the investigated period, did not experience significant land cover changes.
Low-Altitude AVIRIS Data for Mapping Land Cover in Yellowstone National Park: Use of Isodata Clustering Techniques

NASA Technical Reports Server (NTRS)

Spruce, Joseph P.

2001-01-01

Northeast Yellowstone National Park (YNP) has a diversity of forest, range, and wetland cover types. Several remote sensing studies have recently been done in this area, including the NASA Earth Observations Commercial Applications Program (EOCAP) hyperspectral project conducted by Yellowstone Ecosystems Studies (YES) on the use of hyperspectral imaging for assessing riparian and in-stream habitats. In 1999, YES and NASA's Commercial Remote Sensing Program Office began collaborative study of this area, assessing the potential of synergistic use of hyperspectral, synthetic aperture radar (SAR), and multiband thermal data for mapping forest, range, and wetland land cover. Since the beginning, a quality 'reference' land cover map has been desired as a tool for developing and validating other land cover maps produced during the project. This paper recounts an effort to produce such a reference land cover map using low-altitude Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data and unsupervised classification techniques. The main objective of this study is to assess ISODATA classification for mapping land cover in Northeast YNP using select bands of low-altitude AVIRIS data. A secondary, more long-term objective is to assess the potential for improving ISODATA-based classification of land cover through use of principal components analysis and minimum noise fraction (MNF) techniques. This paper will primarily report on work regarding the primary research objective. This study focuses on an AVIRIS cube acquired on July 23, 1999, by the confluence of Soda Butte Creek with the Lamar River. Range and wetland habitats dominate the image with forested habitats being a comparatively minor component of the scene. The scene generally tracks from southwest to northeast. Most of the scene is valley bottom with some lower side slopes occurring on the western portion. Elevations within the AVIRIS scene range from approximately 1998 to 2165 m above sea level, based on US Geological Survey (USGS) 30-m digital elevation model (DEM) data. Despain and the National Park Service (NPS) provide additional description of the study area.
Use of Large-Scale Multi-Configuration EMI Measurements to Characterize Subsurface Structures of the Vadose Zone.

NASA Astrophysics Data System (ADS)

Huisman, J. A.; Brogi, C.; Pätzold, S.; Weihermueller, L.; von Hebel, C.; Van Der Kruk, J.; Vereecken, H.

2017-12-01

Subsurface structures of the vadose zone can play a key role in crop yield potential, especially during water stress periods. Geophysical techniques like electromagnetic induction EMI can provide information about dominant shallow subsurface features. However, previous studies with EMI have typically not reached beyond the field scale. We used high-resolution large-scale multi-configuration EMI measurements to characterize patterns of soil structural organization (layering and texture) and their impact on crop productivity at the km2 scale. We collected EMI data on an agricultural area of 1 km2 (102 ha) near Selhausen (NRW, Germany). The area consists of 51 agricultural fields cropped in rotation. Therefore, measurements were collected between April and December 2016, preferably within few days after the harvest. EMI data were automatically filtered, temperature corrected, and interpolated onto a common grid of 1 m resolution. Inspecting the ECa maps, we identified three main sub-areas with different subsurface heterogeneity. We also identified small-scale geomorphological structures as well as anthropogenic activities such as soil management and buried drainage networks. To identify areas with similar subsurface structures, we applied image classification techniques. We fused ECa maps obtained with different coil distances in a multiband image and applied supervised and unsupervised classification methodologies. Both showed good results in reconstructing observed patterns in plant productivity and the subsurface structures associated with them. However, the supervised methodology proved more efficient in classifying the whole study area. In a second step, we selected hundred locations within the study area and obtained a soil profile description with type, depth, and thickness of the soil horizons. Using this ground truth data it was possible to assign a typical soil profile to each of the main classes obtained from the classification. The proposed methodology was effective in producing a high resolution subsurface model in a large and complex study area that extends well beyond the field scale.
Habitat Mapping and Classification of the Grand Bay National Estuarine Research Reserve using AISA Hyperspectral Imagery

NASA Astrophysics Data System (ADS)

Rose, K.

2012-12-01

Habitat mapping and classification provides essential information for land use planning and ecosystem research, monitoring and management. At the Grand Bay National Estuarine Research Reserve (GRDNERR), Mississippi, habitat characterization of the Grand Bay watershed will also be used to develop a decision-support tool for the NERR's managers and state and local partners. Grand Bay NERR habitat units were identified using a combination of remotely sensed imagery, aerial photography and elevation data. Airborne Imaging Spectrometer for Applications (AISA) hyperspectral data, acquired 5 and 6 May 2010, was analyzed and classified using ENVI v4.8 and v5.0 software. The AISA system was configured to return 63 bands of digital imagery data with a spectral range of 400 to 970 nm (VNIR), spectral resolution (bandwidth) at 8.76 nm, and 1 m spatial resolution. Minimum Noise Fraction (MNF) and Inverse Minimum Noise Fraction were applied to the data prior to using Spectral Angle Mapper ([SAM] supervised) and ISODATA (unsupervised) classification techniques. The resulting class image was exported to ArcGIS 10.0 and visually inspected and compared with the original imagery as well as auxiliary datasets to assist in the attribution of habitat characteristics to the spectral classes, including: National Agricultural Imagery Program (NAIP) aerial photography, Jackson County, MS, 2010; USFWS National Wetlands Inventory, 2007; an existing GRDNERR habitat map (2004), SAV (2009) and salt panne (2002-2003) GIS produced by GRDNERR; and USACE lidar topo-bathymetry, 2005. A field survey to validate the map's accuracy will take place during the 2012 summer season. ENVI's Random Sample generator was used to generate GIS points for a ground-truth survey. The broad range of coastal estuarine habitats and geomorphological features- many of which are transitional and vulnerable to environmental stressors- that have been identified within the GRDNERR point to the value of the Reserve for continued coastal research.
An Embodied Multi-Sensor Fusion Approach to Visual Motion Estimation Using Unsupervised Deep Networks.

PubMed

Shamwell, E Jared; Nothwang, William D; Perlis, Donald

2018-05-04

Aimed at improving size, weight, and power (SWaP)-constrained robotic vision-aided state estimation, we describe our unsupervised, deep convolutional-deconvolutional sensor fusion network, Multi-Hypothesis DeepEfference (MHDE). MHDE learns to intelligently combine noisy heterogeneous sensor data to predict several probable hypotheses for the dense, pixel-level correspondence between a source image and an unseen target image. We show how our multi-hypothesis formulation provides increased robustness against dynamic, heteroscedastic sensor and motion noise by computing hypothesis image mappings and predictions at 76⁻357 Hz depending on the number of hypotheses being generated. MHDE fuses noisy, heterogeneous sensory inputs using two parallel, inter-connected architectural pathways and n (1⁻20 in this work) multi-hypothesis generating sub-pathways to produce n global correspondence estimates between a source and a target image. We evaluated MHDE on the KITTI Odometry dataset and benchmarked it against the vision-only DeepMatching and Deformable Spatial Pyramids algorithms and were able to demonstrate a significant runtime decrease and a performance increase compared to the next-best performing method.
GIS based mapping of land cover changes utilizing multi-temporal remotely sensed image data in Lake Hawassa Watershed, Ethiopia.

PubMed

Nigatu Wondrade; Dick, Øystein B; Tveite, Havard

2014-03-01

Classifying multi-temporal image data to produce thematic maps and quantify land cover changes is one of the most common applications of remote sensing. Mapping land cover changes at the regional level is essential for a wide range of applications including land use planning, decision making, land cover database generation, and as a source of information for sustainable management of natural resources. Land cover changes in Lake Hawassa Watershed, Southern Ethiopia, were investigated using Landsat MSS image data of 1973, and Landsat TM images of 1985, 1995, and 2011, covering a period of nearly four decades. Each image was partitioned in a GIS environment, and classified using an unsupervised algorithm followed by a supervised classification method. A hybrid approach was employed in order to reduce spectral confusion due to high variability of land cover. Classification of satellite image data was performed integrating field data, aerial photographs, topographical maps, medium resolution satellite image (SPOT 20 m), and visual image interpretation. The image data were classified into nine land cover types: water, built-up, cropland, woody vegetation, forest, grassland, swamp, bare land, and scrub. The overall accuracy of the LULC maps ranged from 82.5 to 85.0 %. The achieved accuracies were reasonable, and the observed classification errors were attributable to coarse spatial resolution and pixels containing a mixture of cover types. Land cover change statistics were extracted and tabulated using the ERDAS Imagine software. The results indicated an increase in built-up area, cropland, and bare land areas, and a reduction in the six other land cover classes. Predominant land cover is cropland changing from 43.6 % in 1973 to 56.4 % in 2011. A significant portion of land cover was converted into cropland. Woody vegetation and forest cover which occupied 21.0 and 10.3 % in 1973, respectively, diminished to 13.6 and 5.6 % in 2011. The change in water body was very peculiar in that the area of Lake Hawassa increased from 91.9 km(2) in 1973 to 95.2 km(2) in 2011, while that of Lake Cheleleka whose area was 11.3 km(2) in 1973 totally vanished in 2011 and transformed into mud-flat and grass dominated swamp. The "change and no change" analysis revealed that more than one third (548.0 km(2)) of the total area was exposed to change between 1973 and 2011. This study was useful in identifying the major land cover changes, and the analysis pursued provided a valuable insight into the ongoing changes in the area under investigation.
Analysing land cover and land use change in the Matobo National Park and surroundings in Zimbabwe

NASA Astrophysics Data System (ADS)

Scharsich, Valeska; Mtata, Kupakwashe; Hauhs, Michael; Lange, Holger; Bogner, Christina

2016-04-01

Natural forests are threatened worldwide, therefore their protection in National Parks is essential. Here, we investigate how this protection status affects the land cover. To answer this question, we analyse the surface reflectance of three Landsat images of Matobo National Park and surrounding in Zimbabwe from 1989, 1998 and 2014 to detect changes in land cover in this region. To account for the rolling countryside and the resulting prominent shadows, a topographical correction of the surface reflectance was required. To infer land cover changes it is not only necessary to have some ground data for the current satellite images but also for the old ones. In particular for the older images no recent field study could help to reconstruct these data reliably. In our study we follow the idea that land cover classes of pixels in current images can be transferred to the equivalent pixels of older ones if no changes occurred meanwhile. Therefore we combine unsupervised clustering with supervised classification as follows. At first, we produce a land cover map for 2014. Secondly, we cluster the images with clara, which is similar to k-means, but suitable for large data sets. Whereby the best number of classes were determined to be 4. Thirdly, we locate unchanged pixels with change vector analysis in the images of 1989 and 1998. For these pixels we transfer the corresponding cluster label from 2014 to 1989 and 1998. Subsequently, the classified pixels serve as training data for supervised classification with random forest, which is carried out for each image separately. Finally, we derive land cover classes from the Landsat image in 2014, photographs and Google Earth and transfer them to the other two images. The resulting classes are shrub land; forest/shallow waters; bare soils/fields with some trees/shrubs; and bare light soils/rocks, fields and settlements. Subsequently the three different classifications are compared and land changes are mapped. The main changes are observable in the surroundings of the National Park, especially the common lands have lost their clear boundaries with time. In the National Park, the area of forest increases from 1989 to 2014 from 58% to 61% whereas the area of shrub land decreases by the same amount. The amount of each of the other two classes remains constant. These changes indicate an actual effect of the protection status of the National Park. In our study remote sensing data are the main source to evaluate the effects and the benefits of a protected area without on-side studies. This could be important for regions, where field studies are not possible because of insecure political conditions and only remote sensing data are available.
Filtering large-scale event collections using a combination of supervised and unsupervised learning for event trigger classification.

PubMed

Mehryary, Farrokh; Kaewphan, Suwisa; Hakala, Kai; Ginter, Filip

2016-01-01

Biomedical event extraction is one of the key tasks in biomedical text mining, supporting various applications such as database curation and hypothesis generation. Several systems, some of which have been applied at a large scale, have been introduced to solve this task. Past studies have shown that the identification of the phrases describing biological processes, also known as trigger detection, is a crucial part of event extraction, and notable overall performance gains can be obtained by solely focusing on this sub-task. In this paper we propose a novel approach for filtering falsely identified triggers from large-scale event databases, thus improving the quality of knowledge extraction. Our method relies on state-of-the-art word embeddings, event statistics gathered from the whole biomedical literature, and both supervised and unsupervised machine learning techniques. We focus on EVEX, an event database covering the whole PubMed and PubMed Central Open Access literature containing more than 40 million extracted events. The top most frequent EVEX trigger words are hierarchically clustered, and the resulting cluster tree is pruned to identify words that can never act as triggers regardless of their context. For rarely occurring trigger words we introduce a supervised approach trained on the combination of trigger word classification produced by the unsupervised clustering method and manual annotation. The method is evaluated on the official test set of BioNLP Shared Task on Event Extraction. The evaluation shows that the method can be used to improve the performance of the state-of-the-art event extraction systems. This successful effort also translates into removing 1,338,075 of potentially incorrect events from EVEX, thus greatly improving the quality of the data. The method is not solely bound to the EVEX resource and can be thus used to improve the quality of any event extraction system or database. The data and source code for this work are available at: http://bionlp-www.utu.fi/trigger-clustering/.
Clustering approach for unsupervised segmentation of malarial Plasmodium vivax parasite

NASA Astrophysics Data System (ADS)

Abdul-Nasir, Aimi Salihah; Mashor, Mohd Yusoff; Mohamed, Zeehaida

2017-10-01

Malaria is a global health problem, particularly in Africa and south Asia where it causes countless deaths and morbidity cases. Efficient control and prompt of this disease require early detection and accurate diagnosis due to the large number of cases reported yearly. To achieve this aim, this paper proposes an image segmentation approach via unsupervised pixel segmentation of malaria parasite to automate the diagnosis of malaria. In this study, a modified clustering algorithm namely enhanced k-means (EKM) clustering, is proposed for malaria image segmentation. In the proposed EKM clustering, the concept of variance and a new version of transferring process for clustered members are used to assist the assignation of data to the proper centre during the process of clustering, so that good segmented malaria image can be generated. The effectiveness of the proposed EKM clustering has been analyzed qualitatively and quantitatively by comparing this algorithm with two popular image segmentation techniques namely Otsu's thresholding and k-means clustering. The experimental results show that the proposed EKM clustering has successfully segmented 100 malaria images of P. vivax species with segmentation accuracy, sensitivity and specificity of 99.20%, 87.53% and 99.58%, respectively. Hence, the proposed EKM clustering can be considered as an image segmentation tool for segmenting the malaria images.
An Extended Spectral-Spatial Classification Approach for Hyperspectral Data

NASA Astrophysics Data System (ADS)

Akbari, D.

2017-11-01

In this paper an extended classification approach for hyperspectral imagery based on both spectral and spatial information is proposed. The spatial information is obtained by an enhanced marker-based minimum spanning forest (MSF) algorithm. Three different methods of dimension reduction are first used to obtain the subspace of hyperspectral data: (1) unsupervised feature extraction methods including principal component analysis (PCA), independent component analysis (ICA), and minimum noise fraction (MNF); (2) supervised feature extraction including decision boundary feature extraction (DBFE), discriminate analysis feature extraction (DAFE), and nonparametric weighted feature extraction (NWFE); (3) genetic algorithm (GA). The spectral features obtained are then fed into the enhanced marker-based MSF classification algorithm. In the enhanced MSF algorithm, the markers are extracted from the classification maps obtained by both SVM and watershed segmentation algorithm. To evaluate the proposed approach, the Pavia University hyperspectral data is tested. Experimental results show that the proposed approach using GA achieves an approximately 8 % overall accuracy higher than the original MSF-based algorithm.
Comparison of MSS and TM Data for Landcover Classification in the Chesapeake Bay Area: a Preliminary Report. [Taylor's Island, Maryland

NASA Technical Reports Server (NTRS)

Mulligan, P. J.; Gervin, J. C.; Lu, Y. C.

1985-01-01

An area bordering the Eastern Shore of the Chesapeake Bay was selected for study and classified using unsupervised techniques applied to LANDSAT-2 MSS data and several band combinations of LANDSAT-4 TM data. The accuracies of these Level I land cover classifications were verified using the Taylor's Island USGS 7.5 minute topographic map which was photointerpreted, digitized and rasterized. The the Taylor's Island map, comparing the MSS and TM three band (2 3 4) classifications, the increased resolution of TM produced a small improvement in overall accuracy of 1% correct due primarily to a small improvement, and 1% and 3%, in areas such as water and woodland. This was expected as the MSS data typically produce high accuracies for categories which cover large contiguous areas. However, in the categories covering smaller areas within the map there was generally an improvement of at least 10%. Classification of the important residential category improved 12%, and wetlands were mapped with 11% greater accuracy.
Detection of Coastline Deformation Using Remote Sensing and Geodetic Surveys

NASA Astrophysics Data System (ADS)

Sabuncu, A.; Dogru, A.; Ozener, H.; Turgut, B.

2016-06-01

The coastal areas are being destroyed due to the usage that effect the natural balance. Unconsciously sand mining from the sea for nearshore nourishment and construction uses are the main ones. Physical interferences for mining of sand cause an ecologic threat to the coastal environment. However, use of marine sand is inevitable because of economic reasons or unobtainable land-based sand resources. The most convenient solution in such a protection-usage dilemma is to reduce negative impacts of sand production from marine. This depends on the accurate determination of criteriaon production place, style, and amount of sand. With this motivation, nearshore geodedic surveying studies performed on Kilyos Campus of Bogazici University located on the Black Sea coast, north of Istanbul, Turkey between 2001-2002. The study area extends 1 km in the longshore. Geodetic survey was carried out in the summer of 2001 to detect the initial condition for the shoreline. Long-term seasonal changes in shoreline positions were determined biannually. The coast was measured with post-processed kinematic GPS. Besides, shoreline change has studied using Landsat imagery between the years 1986-2015. The data set of Landsat 5 imageries were dated 05.08.1986 and 31.08.2007 and Landsat 7 imageries were dated 21.07.2001 and 28.07.2015. Landcover types in the study area were analyzed on the basis of pixel based classification method. Firstly, unsupervised classification based on ISODATA (Iterative Self Organizing Data Analysis Technique) has been applied and spectral clusters have been determined that gives prior knowledge about the study area. In the second step, supervised classification was carried out by using the three different approaches which are minimum-distance, parallelepiped and maximum-likelihood. All pixel based classification processes were performed with ENVI 4.8 image processing software. Results of geodetic studies and classification outputs will be presented in this paper.
Scalable High Performance Image Registration Framework by Unsupervised Deep Feature Representations Learning

PubMed Central

Wu, Guorong; Kim, Minjeong; Wang, Qian; Munsell, Brent C.

2015-01-01

Feature selection is a critical step in deformable image registration. In particular, selecting the most discriminative features that accurately and concisely describe complex morphological patterns in image patches improves correspondence detection, which in turn improves image registration accuracy. Furthermore, since more and more imaging modalities are being invented to better identify morphological changes in medical imaging data,, the development of deformable image registration method that scales well to new image modalities or new image applications with little to no human intervention would have a significant impact on the medical image analysis community. To address these concerns, a learning-based image registration framework is proposed that uses deep learning to discover compact and highly discriminative features upon observed imaging data. Specifically, the proposed feature selection method uses a convolutional stacked auto-encoder to identify intrinsic deep feature representations in image patches. Since deep learning is an unsupervised learning method, no ground truth label knowledge is required. This makes the proposed feature selection method more flexible to new imaging modalities since feature representations can be directly learned from the observed imaging data in a very short amount of time. Using the LONI and ADNI imaging datasets, image registration performance was compared to two existing state-of-the-art deformable image registration methods that use handcrafted features. To demonstrate the scalability of the proposed image registration framework image registration experiments were conducted on 7.0-tesla brain MR images. In all experiments, the results showed the new image registration framework consistently demonstrated more accurate registration results when compared to state-of-the-art. PMID:26552069
Scalable High-Performance Image Registration Framework by Unsupervised Deep Feature Representations Learning.

PubMed

Wu, Guorong; Kim, Minjeong; Wang, Qian; Munsell, Brent C; Shen, Dinggang

2016-07-01

Feature selection is a critical step in deformable image registration. In particular, selecting the most discriminative features that accurately and concisely describe complex morphological patterns in image patches improves correspondence detection, which in turn improves image registration accuracy. Furthermore, since more and more imaging modalities are being invented to better identify morphological changes in medical imaging data, the development of deformable image registration method that scales well to new image modalities or new image applications with little to no human intervention would have a significant impact on the medical image analysis community. To address these concerns, a learning-based image registration framework is proposed that uses deep learning to discover compact and highly discriminative features upon observed imaging data. Specifically, the proposed feature selection method uses a convolutional stacked autoencoder to identify intrinsic deep feature representations in image patches. Since deep learning is an unsupervised learning method, no ground truth label knowledge is required. This makes the proposed feature selection method more flexible to new imaging modalities since feature representations can be directly learned from the observed imaging data in a very short amount of time. Using the LONI and ADNI imaging datasets, image registration performance was compared to two existing state-of-the-art deformable image registration methods that use handcrafted features. To demonstrate the scalability of the proposed image registration framework, image registration experiments were conducted on 7.0-T brain MR images. In all experiments, the results showed that the new image registration framework consistently demonstrated more accurate registration results when compared to state of the art.
An unsupervised MVA method to compare specific regions in human breast tumor tissue samples using ToF-SIMS.

PubMed

Bluestein, Blake M; Morrish, Fionnuala; Graham, Daniel J; Guenthoer, Jamie; Hockenbery, David; Porter, Peggy L; Gamble, Lara J

2016-03-21

Imaging time-of-flight secondary ion mass spectrometry (ToF-SIMS) and principal component analysis (PCA) were used to investigate two sets of pre- and post-chemotherapy human breast tumor tissue sections to characterize lipids associated with tumor metabolic flexibility and response to treatment. The micron spatial resolution imaging capability of ToF-SIMS provides a powerful approach to attain spatially-resolved molecular and cellular data from cancerous tissues not available with conventional imaging techniques. Three ca. 1 mm(2) areas per tissue section were analyzed by stitching together 200 μm × 200 μm raster area scans. A method to isolate and analyze specific tissue regions of interest by utilizing PCA of ToF-SIMS images is presented, which allowed separation of cellularized areas from stromal areas. These PCA-generated regions of interest were then used as masks to reconstruct representative spectra from specifically stromal or cellular regions. The advantage of this unsupervised selection method is a reduction in scatter in the spectral PCA results when compared to analyzing all tissue areas or analyzing areas highlighted by a pathologist. Utilizing this method, stromal and cellular regions of breast tissue biopsies taken pre- versus post-chemotherapy demonstrate chemical separation using negatively-charged ion species. In this sample set, the cellular regions were predominantly all cancer cells. Fatty acids (i.e. palmitic, oleic, and stearic), monoacylglycerols, diacylglycerols and vitamin E profiles were distinctively different between the pre- and post-therapy tissues. These results validate a new unsupervised method to isolate and interpret biochemically distinct regions in cancer tissues using imaging ToF-SIMS data. In addition, the method developed here can provide a framework to compare a variety of tissue samples using imaging ToF-SIMS, especially where there is section-to-section variability that makes it difficult to use a serial hematoxylin and eosin (H&E) stained section to direct the SIMS analysis.
High resolution mapping and classification of oyster habitats in nearshore Louisiana using sidescan sonar

USGS Publications Warehouse

Allen, Y.C.; Wilson, C.A.; Roberts, H.H.; Supan, J.

2005-01-01

Sidescan sonar holds great promise as a tool to quantitatively depict the distribution and extent of benthic habitats in Louisiana's turbid estuaries. In this study, we describe an effective protocol for acoustic sampling in this environment. We also compared three methods of classification in detail: mean-based thresholding, supervised, and unsupervised techniques to classify sidescan imagery into categories of mud and shell. Classification results were compared to ground truth results using quadrat and dredge sampling. Supervised classification gave the best overall result (kappa = 75%) when compared to quadrat results. Classification accuracy was less robust when compared to all dredge samples (kappa = 21-56%), but increased greatly (90-100%) when only dredge samples taken from acoustically homogeneous areas were considered. Sidescan sonar when combined with ground truth sampling at an appropriate scale can be effectively used to establish an accurate substrate base map for both research applications and shellfish management. The sidescan imagery presented here also provides, for the first time, a detailed presentation of oyster habitat patchiness and scale in a productive oyster growing area.
Sea ice type dynamics in the Arctic based on Sentinel-1 Data

NASA Astrophysics Data System (ADS)

Babiker, Mohamed; Korosov, Anton; Park, Jeong-Won

2017-04-01

Sea ice observation from satellites has been carried out for more than four decades and is one of the most important applications of EO data in operational monitoring as well as in climate change studies. Several sensors and retrieval methods have been developed and successfully utilized to measure sea ice area, concentration, drift, type, thickness, etc [e.g. Breivik et al., 2009]. Today operational sea ice monitoring and analysis is fully dependent on use of satellite data. However, new and improved satellite systems, such as multi-polarisation Synthetic Apperture Radar (SAR), require further studies to develop more advanced and automated sea ice monitoring methods. In addition, the unprecedented volume of data available from recently launched Sentinel missions provides both challenges and opportunities for studying sea ice dynamics. In this study we investigate sea ice type dynamics in the Fram strait based on Sentinel-1 A, B SAR data. Series of images for the winter season are classified into 4 ice types (young ice, first year ice, multiyear ice and leads) using the new algorithm developed by us for sea ice classification, which is based on segmentation, GLCM calculation, Haralick texture feature extraction, unsupervised and supervised classifications and Support Vector Machine (SVM) [Zakhvatkina et al., 2016; Korosov et al., 2016]. This algorithm is further improved by applying thermal and scalloping noise removal [Park et al. 2016]. Sea ice drift is retrieved from the same series of Sentinel-1 images using the newly developed algorithm based on combination of feature tracking and pattern matching [Mukenhuber et al., 2016]. Time series of these two products (sea ice type and sea ice drift) are combined in order to study sea ice deformation processes at small scales. Zones of sea ice convergence and divergence identified from sea ice drift are compared with ridges and leads identified from texture features. That allows more specific interpretation of SAR imagery and more accurate automatic classification. In addition, the map of four ice types calculated using the texture features from one SAR image is propagated forward using the sea ice drift vectors. The propagated ice type is compared with ice type derived from the next image. The comparison identifies changes in ice type which occurred during drift and allows to reduce uncertainties in sea ice type calculation.
Feature learning and change feature classification based on deep learning for ternary change detection in SAR images

NASA Astrophysics Data System (ADS)

Gong, Maoguo; Yang, Hailun; Zhang, Puzhao

2017-07-01

Ternary change detection aims to detect changes and group the changes into positive change and negative change. It is of great significance in the joint interpretation of spatial-temporal synthetic aperture radar images. In this study, sparse autoencoder, convolutional neural networks (CNN) and unsupervised clustering are combined to solve ternary change detection problem without any supervison. Firstly, sparse autoencoder is used to transform log-ratio difference image into a suitable feature space for extracting key changes and suppressing outliers and noise. And then the learned features are clustered into three classes, which are taken as the pseudo labels for training a CNN model as change feature classifier. The reliable training samples for CNN are selected from the feature maps learned by sparse autoencoder with certain selection rules. Having training samples and the corresponding pseudo labels, the CNN model can be trained by using back propagation with stochastic gradient descent. During its training procedure, CNN is driven to learn the concept of change, and more powerful model is established to distinguish different types of changes. Unlike the traditional methods, the proposed framework integrates the merits of sparse autoencoder and CNN to learn more robust difference representations and the concept of change for ternary change detection. Experimental results on real datasets validate the effectiveness and superiority of the proposed framework.

Visualization of heterogeneity and regional grading of gliomas by multiple features using magnetic resonance-based clustered images.

PubMed

Inano, Rika; Oishi, Naoya; Kunieda, Takeharu; Arakawa, Yoshiki; Kikuchi, Takayuki; Fukuyama, Hidenao; Miyamoto, Susumu

2016-07-26

Preoperative glioma grading is important for therapeutic strategies and influences prognosis. Intratumoral heterogeneity can cause an underestimation of grading because of the sampling error in biopsies. We developed a voxel-based unsupervised clustering method with multiple magnetic resonance imaging (MRI)-derived features using a self-organizing map followed by K-means. This method produced novel magnetic resonance-based clustered images (MRcIs) that enabled the visualization of glioma grades in 36 patients. The 12-class MRcIs revealed the highest classification performance for the prediction of glioma grading (area under the receiver operating characteristic curve = 0.928; 95% confidential interval = 0.920-0.936). Furthermore, we also created 12-class MRcIs in four new patients using the previous data from the 36 patients as training data and obtained tissue sections of the classes 11 and 12, which were significantly higher in high-grade gliomas (HGGs), and those of classes 4, 5 and 9, which were not significantly different between HGGs and low-grade gliomas (LGGs), according to a MRcI-based navigational system. The tissues of classes 11 and 12 showed features of malignant glioma, whereas those of classes 4, 5 and 9 showed LGGs without anaplastic features. These results suggest that the proposed voxel-based clustering method provides new insights into preoperative regional glioma grading.
Sparse alignment for robust tensor learning.

PubMed

Lai, Zhihui; Wong, Wai Keung; Xu, Yong; Zhao, Cairong; Sun, Mingming

2014-10-01

Multilinear/tensor extensions of manifold learning based algorithms have been widely used in computer vision and pattern recognition. This paper first provides a systematic analysis of the multilinear extensions for the most popular methods by using alignment techniques, thereby obtaining a general tensor alignment framework. From this framework, it is easy to show that the manifold learning based tensor learning methods are intrinsically different from the alignment techniques. Based on the alignment framework, a robust tensor learning method called sparse tensor alignment (STA) is then proposed for unsupervised tensor feature extraction. Different from the existing tensor learning methods, L1- and L2-norms are introduced to enhance the robustness in the alignment step of the STA. The advantage of the proposed technique is that the difficulty in selecting the size of the local neighborhood can be avoided in the manifold learning based tensor feature extraction algorithms. Although STA is an unsupervised learning method, the sparsity encodes the discriminative information in the alignment step and provides the robustness of STA. Extensive experiments on the well-known image databases as well as action and hand gesture databases by encoding object images as tensors demonstrate that the proposed STA algorithm gives the most competitive performance when compared with the tensor-based unsupervised learning methods.
Surface defect detection in tiling Industries using digital image processing methods: analysis and evaluation.

PubMed

Karimi, Mohammad H; Asemani, Davud

2014-05-01

Ceramic and tile industries should indispensably include a grading stage to quantify the quality of products. Actually, human control systems are often used for grading purposes. An automatic grading system is essential to enhance the quality control and marketing of the products. Since there generally exist six different types of defects originating from various stages of tile manufacturing lines with distinct textures and morphologies, many image processing techniques have been proposed for defect detection. In this paper, a survey has been made on the pattern recognition and image processing algorithms which have been used to detect surface defects. Each method appears to be limited for detecting some subgroup of defects. The detection techniques may be divided into three main groups: statistical pattern recognition, feature vector extraction and texture/image classification. The methods such as wavelet transform, filtering, morphology and contourlet transform are more effective for pre-processing tasks. Others including statistical methods, neural networks and model-based algorithms can be applied to extract the surface defects. Although, statistical methods are often appropriate for identification of large defects such as Spots, but techniques such as wavelet processing provide an acceptable response for detection of small defects such as Pinhole. A thorough survey is made in this paper on the existing algorithms in each subgroup. Also, the evaluation parameters are discussed including supervised and unsupervised parameters. Using various performance parameters, different defect detection algorithms are compared and evaluated. Copyright © 2013 ISA. Published by Elsevier Ltd. All rights reserved.
Model-based classification of CPT data and automated lithostratigraphic mapping for high-resolution characterization of a heterogeneous sedimentary aquifer

PubMed Central

Mallants, Dirk; Batelaan, Okke; Gedeon, Matej; Huysmans, Marijke; Dassargues, Alain

2017-01-01

Cone penetration testing (CPT) is one of the most efficient and versatile methods currently available for geotechnical, lithostratigraphic and hydrogeological site characterization. Currently available methods for soil behaviour type classification (SBT) of CPT data however have severe limitations, often restricting their application to a local scale. For parameterization of regional groundwater flow or geotechnical models, and delineation of regional hydro- or lithostratigraphy, regional SBT classification would be very useful. This paper investigates the use of model-based clustering for SBT classification, and the influence of different clustering approaches on the properties and spatial distribution of the obtained soil classes. We additionally propose a methodology for automated lithostratigraphic mapping of regionally occurring sedimentary units using SBT classification. The methodology is applied to a large CPT dataset, covering a groundwater basin of ~60 km2 with predominantly unconsolidated sandy sediments in northern Belgium. Results show that the model-based approach is superior in detecting the true lithological classes when compared to more frequently applied unsupervised classification approaches or literature classification diagrams. We demonstrate that automated mapping of lithostratigraphic units using advanced SBT classification techniques can provide a large gain in efficiency, compared to more time-consuming manual approaches and yields at least equally accurate results. PMID:28467468
Model-based classification of CPT data and automated lithostratigraphic mapping for high-resolution characterization of a heterogeneous sedimentary aquifer.

PubMed

Rogiers, Bart; Mallants, Dirk; Batelaan, Okke; Gedeon, Matej; Huysmans, Marijke; Dassargues, Alain

2017-01-01

Cone penetration testing (CPT) is one of the most efficient and versatile methods currently available for geotechnical, lithostratigraphic and hydrogeological site characterization. Currently available methods for soil behaviour type classification (SBT) of CPT data however have severe limitations, often restricting their application to a local scale. For parameterization of regional groundwater flow or geotechnical models, and delineation of regional hydro- or lithostratigraphy, regional SBT classification would be very useful. This paper investigates the use of model-based clustering for SBT classification, and the influence of different clustering approaches on the properties and spatial distribution of the obtained soil classes. We additionally propose a methodology for automated lithostratigraphic mapping of regionally occurring sedimentary units using SBT classification. The methodology is applied to a large CPT dataset, covering a groundwater basin of ~60 km2 with predominantly unconsolidated sandy sediments in northern Belgium. Results show that the model-based approach is superior in detecting the true lithological classes when compared to more frequently applied unsupervised classification approaches or literature classification diagrams. We demonstrate that automated mapping of lithostratigraphic units using advanced SBT classification techniques can provide a large gain in efficiency, compared to more time-consuming manual approaches and yields at least equally accurate results.
Correlation between aircraft MSS and LIDAR remotely sensed data on a forested wetland in South Carolina

NASA Technical Reports Server (NTRS)

Jensen, John R.; Hodgson, Michael E.; Mackey, Halkard E., Jr.; Krabill, William

1987-01-01

Wetlands in a portion of the Savannah River swamp forest, the Steel Creek Delta, were mapped using April 26, 1985 high-resolution aircraft multispectral scanner (MSS) data. Due to the complex spectral characteristics of the wetland vegetation, it was necessary to implement several techniques in the classification of the MSS imagery of the Steel Creek Delta. In particular, when performing unsupervised classification, an iterative cluster busting technique was used which simplified the cluster labeling process. In addition to the MSS data, light detecting and ranging (LIDAR) data were acquired by National Aeronautics and Space Administration (NASA) personnel along two flightlines over the Steel Creek Delta. These data were registered with the wetland classification map and correlated. Statistical analyses demonstrated that the laser derived canopy height information was significantly correlated with the Steel Creek Delta wetland classes encountered along the profiling transect of the LIDAR data.
Clonal Selection Based Artificial Immune System for Generalized Pattern Recognition

NASA Technical Reports Server (NTRS)

Huntsberger, Terry

2011-01-01

The last two decades has seen a rapid increase in the application of AIS (Artificial Immune Systems) modeled after the human immune system to a wide range of areas including network intrusion detection, job shop scheduling, classification, pattern recognition, and robot control. JPL (Jet Propulsion Laboratory) has developed an integrated pattern recognition/classification system called AISLE (Artificial Immune System for Learning and Exploration) based on biologically inspired models of B-cell dynamics in the immune system. When used for unsupervised or supervised classification, the method scales linearly with the number of dimensions, has performance that is relatively independent of the total size of the dataset, and has been shown to perform as well as traditional clustering methods. When used for pattern recognition, the method efficiently isolates the appropriate matches in the data set. The paper presents the underlying structure of AISLE and the results from a number of experimental studies.
Transient classification in LIGO data using difference boosting neural network

NASA Astrophysics Data System (ADS)

Mukund, N.; Abraham, S.; Kandhasamy, S.; Mitra, S.; Philip, N. S.

2017-05-01

Detection and classification of transients in data from gravitational wave detectors are crucial for efficient searches for true astrophysical events and identification of noise sources. We present a hybrid method for classification of short duration transients seen in gravitational wave data using both supervised and unsupervised machine learning techniques. To train the classifiers, we use the relative wavelet energy and the corresponding entropy obtained by applying one-dimensional wavelet decomposition on the data. The prediction accuracy of the trained classifier on nine simulated classes of gravitational wave transients and also LIGO's sixth science run hardware injections are reported. Targeted searches for a couple of known classes of nonastrophysical signals in the first observational run of Advanced LIGO data are also presented. The ability to accurately identify transient classes using minimal training samples makes the proposed method a useful tool for LIGO detector characterization as well as searches for short duration gravitational wave signals.
Random forests for classification in ecology

USGS Publications Warehouse

Cutler, D.R.; Edwards, T.C.; Beard, K.H.; Cutler, A.; Hess, K.T.; Gibson, J.; Lawler, J.J.

2007-01-01

Classification procedures are some of the most widely used statistical methods in ecology. Random forests (RF) is a new and powerful statistical classifier that is well established in other disciplines but is relatively unknown in ecology. Advantages of RF compared to other statistical classifiers include (1) very high classification accuracy; (2) a novel method of determining variable importance; (3) ability to model complex interactions among predictor variables; (4) flexibility to perform several types of statistical data analysis, including regression, classification, survival analysis, and unsupervised learning; and (5) an algorithm for imputing missing values. We compared the accuracies of RF and four other commonly used statistical classifiers using data on invasive plant species presence in Lava Beds National Monument, California, USA, rare lichen species presence in the Pacific Northwest, USA, and nest sites for cavity nesting birds in the Uinta Mountains, Utah, USA. We observed high classification accuracy in all applications as measured by cross-validation and, in the case of the lichen data, by independent test data, when comparing RF to other common classification methods. We also observed that the variables that RF identified as most important for classifying invasive plant species coincided with expectations based on the literature. ?? 2007 by the Ecological Society of America.
Image segmentation using fuzzy LVQ clustering networks

NASA Technical Reports Server (NTRS)

Tsao, Eric Chen-Kuo; Bezdek, James C.; Pal, Nikhil R.

1992-01-01

In this note we formulate image segmentation as a clustering problem. Feature vectors extracted from a raw image are clustered into subregions, thereby segmenting the image. A fuzzy generalization of a Kohonen learning vector quantization (LVQ) which integrates the Fuzzy c-Means (FCM) model with the learning rate and updating strategies of the LVQ is used for this task. This network, which segments images in an unsupervised manner, is thus related to the FCM optimization problem. Numerical examples on photographic and magnetic resonance images are given to illustrate this approach to image segmentation.
Automatic Temporal Tracking of Supra-Glacial Lakes

NASA Astrophysics Data System (ADS)

Liang, Y.; Lv, Q.; Gallaher, D. W.; Fanning, D.

2010-12-01

During the recent years, supra-glacial lakes in Greenland have attracted extensive global attention as they potentially play an important role in glacier movement, sea level rise, and climate change. Previous works focused on classification methods and individual cloud-free satellite images, which have limited capabilities in terms of tracking changes of lakes over time. The challenges of tracking supra-glacial lakes automatically include (1) massive amount of satellite images with diverse qualities and frequent cloud coverage, and (2) diversity and dynamics of large number of supra-glacial lakes on the Greenland ice sheet. In this study, we develop an innovative method to automatically track supra-glacial lakes temporally using the Moderate Resolution Imaging Spectroradiometer (MODIS) time-series data. The method works for both cloudy and cloud-free data and is unsupervised, i.e., no manual identification is required. After selecting the highest-quality image within each time interval, our method automatically detects supra-glacial lakes in individual images, using adaptive thresholding to handle diverse image qualities. We then track lakes across time series of images as lakes appear, change in size, and disappear. Using multi-year MODIS data during melting season, we demonstrate that this new method can detect and track supra-glacial lakes in both space and time with 95% accuracy. Attached figure shows an example of the current result. Detailed analysis of the temporal variation of detected lakes will be presented. (a) One of our experimental data. The Investigated region is centered at Jakobshavn Isbrae glacier in west Greenland. (b) Enlarged view of part of ice sheet. It is partially cloudy and with supra-glacial lakes on it. Lakes are shown as dark spots. (c) Current result. Red spots are detected lakes.
An Unsupervised Change Detection Method Using Time-Series of PolSAR Images from Radarsat-2 and GaoFen-3.

PubMed

Liu, Wensong; Yang, Jie; Zhao, Jinqi; Shi, Hongtao; Yang, Le

2018-02-12

The traditional unsupervised change detection methods based on the pixel level can only detect the changes between two different times with same sensor, and the results are easily affected by speckle noise. In this paper, a novel method is proposed to detect change based on time-series data from different sensors. Firstly, the overall difference image of the time-series PolSAR is calculated by omnibus test statistics, and difference images between any two images in different times are acquired by R j test statistics. Secondly, the difference images are segmented with a Generalized Statistical Region Merging (GSRM) algorithm which can suppress the effect of speckle noise. Generalized Gaussian Mixture Model (GGMM) is then used to obtain the time-series change detection maps in the final step of the proposed method. To verify the effectiveness of the proposed method, we carried out the experiment of change detection using time-series PolSAR images acquired by Radarsat-2 and Gaofen-3 over the city of Wuhan, in China. Results show that the proposed method can not only detect the time-series change from different sensors, but it can also better suppress the influence of speckle noise and improve the overall accuracy and Kappa coefficient.
Deep learning-based subdivision approach for large scale macromolecules structure recovery from electron cryo tomograms.

PubMed

Xu, Min; Chai, Xiaoqi; Muthakana, Hariank; Liang, Xiaodan; Yang, Ge; Zeev-Ben-Mordehai, Tzviya; Xing, Eric P

2017-07-15

Cellular Electron CryoTomography (CECT) enables 3D visualization of cellular organization at near-native state and in sub-molecular resolution, making it a powerful tool for analyzing structures of macromolecular complexes and their spatial organizations inside single cells. However, high degree of structural complexity together with practical imaging limitations makes the systematic de novo discovery of structures within cells challenging. It would likely require averaging and classifying millions of subtomograms potentially containing hundreds of highly heterogeneous structural classes. Although it is no longer difficult to acquire CECT data containing such amount of subtomograms due to advances in data acquisition automation, existing computational approaches have very limited scalability or discrimination ability, making them incapable of processing such amount of data. To complement existing approaches, in this article we propose a new approach for subdividing subtomograms into smaller but relatively homogeneous subsets. The structures in these subsets can then be separately recovered using existing computation intensive methods. Our approach is based on supervised structural feature extraction using deep learning, in combination with unsupervised clustering and reference-free classification. Our experiments show that, compared with existing unsupervised rotation invariant feature and pose-normalization based approaches, our new approach achieves significant improvements in both discrimination ability and scalability. More importantly, our new approach is able to discover new structural classes and recover structures that do not exist in training data. Source code freely available at http://www.cs.cmu.edu/∼mxu1/software . mxu1@cs.cmu.edu. Supplementary data are available at Bioinformatics online. © The Author 2017. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com
BlobContours: adapting Blobworld for supervised color- and texture-based image segmentation

NASA Astrophysics Data System (ADS)

Vogel, Thomas; Nguyen, Dinh Quyen; Dittmann, Jana

2006-01-01

Extracting features is the first and one of the most crucial steps in recent image retrieval process. While the color features and the texture features of digital images can be extracted rather easily, the shape features and the layout features depend on reliable image segmentation. Unsupervised image segmentation, often used in image analysis, works on merely syntactical basis. That is, what an unsupervised segmentation algorithm can segment is only regions, but not objects. To obtain high-level objects, which is desirable in image retrieval, human assistance is needed. Supervised image segmentations schemes can improve the reliability of segmentation and segmentation refinement. In this paper we propose a novel interactive image segmentation technique that combines the reliability of a human expert with the precision of automated image segmentation. The iterative procedure can be considered a variation on the Blobworld algorithm introduced by Carson et al. from EECS Department, University of California, Berkeley. Starting with an initial segmentation as provided by the Blobworld framework, our algorithm, namely BlobContours, gradually updates it by recalculating every blob, based on the original features and the updated number of Gaussians. Since the original algorithm has hardly been designed for interactive processing we had to consider additional requirements for realizing a supervised segmentation scheme on the basis of Blobworld. Increasing transparency of the algorithm by applying usercontrolled iterative segmentation, providing different types of visualization for displaying the segmented image and decreasing computational time of segmentation are three major requirements which are discussed in detail.
Unsupervised fuzzy segmentation of 3D magnetic resonance brain images

NASA Astrophysics Data System (ADS)

Velthuizen, Robert P.; Hall, Lawrence O.; Clarke, Laurence P.; Bensaid, Amine M.; Arrington, J. A.; Silbiger, Martin L.

1993-07-01

Unsupervised fuzzy methods are proposed for segmentation of 3D Magnetic Resonance images of the brain. Fuzzy c-means (FCM) has shown promising results for segmentation of single slices. FCM has been investigated for volume segmentations, both by combining results of single slices and by segmenting the full volume. Different strategies and initializations have been tried. In particular, two approaches have been used: (1) a method by which, iteratively, the furthest sample is split off to form a new cluster center, and (2) the traditional FCM in which the membership grade matrix is initialized in some way. Results have been compared with volume segmentations by k-means and with two supervised methods, k-nearest neighbors and region growing. Results of individual segmentations are presented as well as comparisons on the application of the different methods to a number of tumor patient data sets.
Naive scoring of human sleep based on a hidden Markov model of the electroencephalogram.

PubMed

Yaghouby, Farid; Modur, Pradeep; Sunderam, Sridhar

2014-01-01

Clinical sleep scoring involves tedious visual review of overnight polysomnograms by a human expert. Many attempts have been made to automate the process by training computer algorithms such as support vector machines and hidden Markov models (HMMs) to replicate human scoring. Such supervised classifiers are typically trained on scored data and then validated on scored out-of-sample data. Here we describe a methodology based on HMMs for scoring an overnight sleep recording without the benefit of a trained initial model. The number of states in the data is not known a priori and is optimized using a Bayes information criterion. When tested on a 22-subject database, this unsupervised classifier agreed well with human scores (mean of Cohen's kappa > 0.7). The HMM also outperformed other unsupervised classifiers (Gaussian mixture models, k-means, and linkage trees), that are capable of naive classification but do not model dynamics, by a significant margin (p < 0.05).
Robust Joint Graph Sparse Coding for Unsupervised Spectral Feature Selection.

PubMed

Zhu, Xiaofeng; Li, Xuelong; Zhang, Shichao; Ju, Chunhua; Wu, Xindong

2017-06-01

In this paper, we propose a new unsupervised spectral feature selection model by embedding a graph regularizer into the framework of joint sparse regression for preserving the local structures of data. To do this, we first extract the bases of training data by previous dictionary learning methods and, then, map original data into the basis space to generate their new representations, by proposing a novel joint graph sparse coding (JGSC) model. In JGSC, we first formulate its objective function by simultaneously taking subspace learning and joint sparse regression into account, then, design a new optimization solution to solve the resulting objective function, and further prove the convergence of the proposed solution. Furthermore, we extend JGSC to a robust JGSC (RJGSC) via replacing the least square loss function with a robust loss function, for achieving the same goals and also avoiding the impact of outliers. Finally, experimental results on real data sets showed that both JGSC and RJGSC outperformed the state-of-the-art algorithms in terms of k -nearest neighbor classification performance.
Self-Organizing Hidden Markov Model Map (SOHMMM).

PubMed

Ferles, Christos; Stafylopatis, Andreas

2013-12-01

A hybrid approach combining the Self-Organizing Map (SOM) and the Hidden Markov Model (HMM) is presented. The Self-Organizing Hidden Markov Model Map (SOHMMM) establishes a cross-section between the theoretic foundations and algorithmic realizations of its constituents. The respective architectures and learning methodologies are fused in an attempt to meet the increasing requirements imposed by the properties of deoxyribonucleic acid (DNA), ribonucleic acid (RNA), and protein chain molecules. The fusion and synergy of the SOM unsupervised training and the HMM dynamic programming algorithms bring forth a novel on-line gradient descent unsupervised learning algorithm, which is fully integrated into the SOHMMM. Since the SOHMMM carries out probabilistic sequence analysis with little or no prior knowledge, it can have a variety of applications in clustering, dimensionality reduction and visualization of large-scale sequence spaces, and also, in sequence discrimination, search and classification. Two series of experiments based on artificial sequence data and splice junction gene sequences demonstrate the SOHMMM's characteristics and capabilities. Copyright © 2013 Elsevier Ltd. All rights reserved.
Examining change detection approaches for tropical mangrove monitoring

USGS Publications Warehouse

Myint, Soe W.; Franklin, Janet; Buenemann, Michaela; Kim, Won; Giri, Chandra

2014-01-01

This study evaluated the effectiveness of different band combinations and classifiers (unsupervised, supervised, object-oriented nearest neighbor, and object-oriented decision rule) for quantifying mangrove forest change using multitemporal Landsat data. A discriminant analysis using spectra of different vegetation types determined that bands 2 (0.52 to 0.6 μm), 5 (1.55 to 1.75 μm), and 7 (2.08 to 2.35 μm) were the most effective bands for differentiating mangrove forests from surrounding land cover types. A ranking of thirty-six change maps, produced by comparing the classification accuracy of twelve change detection approaches, was used. The object-based Nearest Neighbor classifier produced the highest mean overall accuracy (84 percent) regardless of band combinations. The automated decision rule-based approach (mean overall accuracy of 88 percent) as well as a composite of bands 2, 5, and 7 used with the unsupervised classifier and the same composite or all band difference with the object-oriented Nearest Neighbor classifier were the most effective approaches.
The Convallis Rule for Unsupervised Learning in Cortical Networks

PubMed Central

Yger, Pierre; Harris, Kenneth D.

2013-01-01

The phenomenology and cellular mechanisms of cortical synaptic plasticity are becoming known in increasing detail, but the computational principles by which cortical plasticity enables the development of sensory representations are unclear. Here we describe a framework for cortical synaptic plasticity termed the “Convallis rule”, mathematically derived from a principle of unsupervised learning via constrained optimization. Implementation of the rule caused a recurrent cortex-like network of simulated spiking neurons to develop rate representations of real-world speech stimuli, enabling classification by a downstream linear decoder. Applied to spike patterns used in in vitro plasticity experiments, the rule reproduced multiple results including and beyond STDP. However STDP alone produced poorer learning performance. The mathematical form of the rule is consistent with a dual coincidence detector mechanism that has been suggested by experiments in several synaptic classes of juvenile neocortex. Based on this confluence of normative, phenomenological, and mechanistic evidence, we suggest that the rule may approximate a fundamental computational principle of the neocortex. PMID:24204224

Some links on this page may take you to non-federal websites. Their policies may differ from this site.