Classification images for localization performance in ramp-spectrum noise.

PubMed

Abbey, Craig K; Samuelson, Frank W; Zeng, Rongping; Boone, John M; Eckstein, Miguel P; Myers, Kyle

2018-05-01

This study investigates forced localization of targets in simulated images with statistical properties similar to trans-axial sections of x-ray computed tomography (CT) volumes. A total of 24 imaging conditions are considered, comprising two target sizes, three levels of background variability, and four levels of frequency apodization. The goal of the study is to better understand how human observers perform forced-localization tasks in images with CT-like statistical properties. The transfer properties of CT systems are modeled by a shift-invariant transfer function in addition to apodization filters that modulate high spatial frequencies. The images contain noise that is the combination of a ramp-spectrum component, simulating the effect of acquisition noise in CT, and a power-law component, simulating the effect of normal anatomy in the background, which are modulated by the apodization filter as well. Observer performance is characterized using two psychophysical techniques: efficiency analysis and classification image analysis. Observer efficiency quantifies how much diagnostic information is being used by observers to perform a task, and classification images show how that information is being accessed in the form of a perceptual filter. Psychophysical studies from five subjects form the basis of the results. Observer efficiency ranges from 29% to 77% across the different conditions. The lowest efficiency is observed in conditions with uniform backgrounds, where significant effects of apodization are found. The classification images, estimated using smoothing windows, suggest that human observers use center-surround filters to perform the task, and these are subjected to a number of subsequent analyses. When implemented as a scanning linear filter, the classification images appear to capture most of the observer variability in efficiency (r 2 = 0.86). The frequency spectra of the classification images show that frequency weights generally appear bandpass in nature, with peak frequency and bandwidth that vary with statistical properties of the images. In these experiments, the classification images appear to capture important features of human-observer performance. Frequency apodization only appears to have a significant effect on performance in the absence of anatomical variability, where the observers appear to underweight low spatial frequencies that have relatively little noise. Frequency weights derived from the classification images generally have a bandpass structure, with adaptation to different conditions seen in the peak frequency and bandwidth. The classification image spectra show relatively modest changes in response to different levels of apodization, with some evidence that observers are attempting to rebalance the apodized spectrum presented to them. © 2018 American Association of Physicists in Medicine.

Skin Lesion Analysis towards Melanoma Detection Using Deep Learning Network.

PubMed

Li, Yuexiang; Shen, Linlin

2018-02-11

Skin lesions are a severe disease globally. Early detection of melanoma in dermoscopy images significantly increases the survival rate. However, the accurate recognition of melanoma is extremely challenging due to the following reasons: low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc. Hence, reliable automatic detection of skin tumors is very useful to increase the accuracy and efficiency of pathologists. In this paper, we proposed two deep learning methods to address three main tasks emerging in the area of skin lesion image processing, i.e., lesion segmentation (task 1), lesion dermoscopic feature extraction (task 2) and lesion classification (task 3). A deep learning framework consisting of two fully convolutional residual networks (FCRN) is proposed to simultaneously produce the segmentation result and the coarse classification result. A lesion index calculation unit (LICU) is developed to refine the coarse classification results by calculating the distance heat-map. A straight-forward CNN is proposed for the dermoscopic feature extraction task. The proposed deep learning frameworks were evaluated on the ISIC 2017 dataset. Experimental results show the promising accuracies of our frameworks, i.e., 0.753 for task 1, 0.848 for task 2 and 0.912 for task 3 were achieved.

Event-Related fMRI of Category Learning: Differences in Classification and Feedback Networks

ERIC Educational Resources Information Center

Little, Deborah M.; Shin, Silvia S.; Sisco, Shannon M.; Thulborn, Keith R.

2006-01-01

Eighteen healthy young adults underwent event-related (ER) functional magnetic resonance imaging (fMRI) of the brain while performing a visual category learning task. The specific category learning task required subjects to extract the rules that guide classification of quasi-random patterns of dots into categories. Following each classification…

Early differential processing of material images: Evidence from ERP classification.

PubMed

Wiebel, Christiane B; Valsecchi, Matteo; Gegenfurtner, Karl R

2014-06-24

Investigating the temporal dynamics of natural image processing using event-related potentials (ERPs) has a long tradition in object recognition research. In a classical Go-NoGo task two characteristic effects have been emphasized: an early task independent category effect and a later task-dependent target effect. Here, we set out to use this well-established Go-NoGo paradigm to study the time course of material categorization. Material perception has gained more and more interest over the years as its importance in natural viewing conditions has been ignored for a long time. In addition to analyzing standard ERPs, we conducted a single trial ERP pattern analysis. To validate this procedure, we also measured ERPs in two object categories (people and animals). Our linear classification procedure was able to largely capture the overall pattern of results from the canonical analysis of the ERPs and even extend it. We replicate the known target effect (differential Go-NoGo potential at frontal sites) for the material images. Furthermore, we observe task-independent differential activity between the two material categories as early as 140 ms after stimulus onset. Using our linear classification approach, we show that material categories can be differentiated consistently based on the ERP pattern in single trials around 100 ms after stimulus onset, independent of the target-related status. This strengthens the idea of early differential visual processing of material categories independent of the task, probably due to differences in low-level image properties and suggests pattern classification of ERP topographies as a strong instrument for investigating electrophysiological brain activity. © 2014 ARVO.

Video based object representation and classification using multiple covariance matrices.

PubMed

Zhang, Yurong; Liu, Quan

2017-01-01

Video based object recognition and classification has been widely studied in computer vision and image processing area. One main issue of this task is to develop an effective representation for video. This problem can generally be formulated as image set representation. In this paper, we present a new method called Multiple Covariance Discriminative Learning (MCDL) for image set representation and classification problem. The core idea of MCDL is to represent an image set using multiple covariance matrices with each covariance matrix representing one cluster of images. Firstly, we use the Nonnegative Matrix Factorization (NMF) method to do image clustering within each image set, and then adopt Covariance Discriminative Learning on each cluster (subset) of images. At last, we adopt KLDA and nearest neighborhood classification method for image set classification. Promising experimental results on several datasets show the effectiveness of our MCDL method.

Label-aligned Multi-task Feature Learning for Multimodal Classification of Alzheimer’s Disease and Mild Cognitive Impairment

PubMed Central

Zu, Chen; Jie, Biao; Liu, Mingxia; Chen, Songcan

2015-01-01

Multimodal classification methods using different modalities of imaging and non-imaging data have recently shown great advantages over traditional single-modality-based ones for diagnosis and prognosis of Alzheimer’s disease (AD), as well as its prodromal stage, i.e., mild cognitive impairment (MCI). However, to the best of our knowledge, most existing methods focus on mining the relationship across multiple modalities of the same subjects, while ignoring the potentially useful relationship across different subjects. Accordingly, in this paper, we propose a novel learning method for multimodal classification of AD/MCI, by fully exploring the relationships across both modalities and subjects. Specifically, our proposed method includes two subsequent components, i.e., label-aligned multi-task feature selection and multimodal classification. In the first step, the feature selection learning from multiple modalities are treated as different learning tasks and a group sparsity regularizer is imposed to jointly select a subset of relevant features. Furthermore, to utilize the discriminative information among labeled subjects, a new label-aligned regularization term is added into the objective function of standard multi-task feature selection, where label-alignment means that all multi-modality subjects with the same class labels should be closer in the new feature-reduced space. In the second step, a multi-kernel support vector machine (SVM) is adopted to fuse the selected features from multi-modality data for final classification. To validate our method, we perform experiments on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database using baseline MRI and FDG-PET imaging data. The experimental results demonstrate that our proposed method achieves better classification performance compared with several state-of-the-art methods for multimodal classification of AD/MCI. PMID:26572145

Image Processing and Computer Aided Diagnosis in Computed Tomography of the Breast

DTIC Science & Technology

2007-03-01

TERMS breast imaging, breast CT, scatter compensation, denoising, CAD , Cone-beam CT 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF...clinical projection images. The CAD tool based on signal known exactly (SKE) scenario is under development. Task 6: Test and compare the...performances of the CAD developed in Task 5 applied to processed projection data from Task 1 with the CAD performance on the projection data without Bayesian

Spatial-spectral blood cell classification with microscopic hyperspectral imagery

NASA Astrophysics Data System (ADS)

Ran, Qiong; Chang, Lan; Li, Wei; Xu, Xiaofeng

2017-10-01

Microscopic hyperspectral images provide a new way for blood cell examination. The hyperspectral imagery can greatly facilitate the classification of different blood cells. In this paper, the microscopic hyperspectral images are acquired by connecting the microscope and the hyperspectral imager, and then tested for blood cell classification. For combined use of the spectral and spatial information provided by hyperspectral images, a spatial-spectral classification method is improved from the classical extreme learning machine (ELM) by integrating spatial context into the image classification task with Markov random field (MRF) model. Comparisons are done among ELM, ELM-MRF, support vector machines(SVM) and SVMMRF methods. Results show the spatial-spectral classification methods(ELM-MRF, SVM-MRF) perform better than pixel-based methods(ELM, SVM), and the proposed ELM-MRF has higher precision and show more accurate location of cells.

Classification of multiple sclerosis lesions using adaptive dictionary learning.

PubMed

Deshpande, Hrishikesh; Maurel, Pierre; Barillot, Christian

2015-12-01

This paper presents a sparse representation and an adaptive dictionary learning based method for automated classification of multiple sclerosis (MS) lesions in magnetic resonance (MR) images. Manual delineation of MS lesions is a time-consuming task, requiring neuroradiology experts to analyze huge volume of MR data. This, in addition to the high intra- and inter-observer variability necessitates the requirement of automated MS lesion classification methods. Among many image representation models and classification methods that can be used for such purpose, we investigate the use of sparse modeling. In the recent years, sparse representation has evolved as a tool in modeling data using a few basis elements of an over-complete dictionary and has found applications in many image processing tasks including classification. We propose a supervised classification approach by learning dictionaries specific to the lesions and individual healthy brain tissues, which include white matter (WM), gray matter (GM) and cerebrospinal fluid (CSF). The size of the dictionaries learned for each class plays a major role in data representation but it is an even more crucial element in the case of competitive classification. Our approach adapts the size of the dictionary for each class, depending on the complexity of the underlying data. The algorithm is validated using 52 multi-sequence MR images acquired from 13 MS patients. The results demonstrate the effectiveness of our approach in MS lesion classification. Copyright © 2015 Elsevier Ltd. All rights reserved.

Hyperspectral image classification based on local binary patterns and PCANet

NASA Astrophysics Data System (ADS)

Yang, Huizhen; Gao, Feng; Dong, Junyu; Yang, Yang

2018-04-01

Hyperspectral image classification has been well acknowledged as one of the challenging tasks of hyperspectral data processing. In this paper, we propose a novel hyperspectral image classification framework based on local binary pattern (LBP) features and PCANet. In the proposed method, linear prediction error (LPE) is first employed to select a subset of informative bands, and LBP is utilized to extract texture features. Then, spectral and texture features are stacked into a high dimensional vectors. Next, the extracted features of a specified position are transformed to a 2-D image. The obtained images of all pixels are fed into PCANet for classification. Experimental results on real hyperspectral dataset demonstrate the effectiveness of the proposed method.

Creating a classification of image types in the medical literature for visual categorization

NASA Astrophysics Data System (ADS)

Müller, Henning; Kalpathy-Cramer, Jayashree; Demner-Fushman, Dina; Antani, Sameer

2012-02-01

Content-based image retrieval (CBIR) from specialized collections has often been proposed for use in such areas as diagnostic aid, clinical decision support, and teaching. The visual retrieval from broad image collections such as teaching files, the medical literature or web images, by contrast, has not yet reached a high maturity level compared to textual information retrieval. Visual image classification into a relatively small number of classes (20-100) on the other hand, has shown to deliver good results in several benchmarks. It is, however, currently underused as a basic technology for retrieval tasks, for example, to limit the search space. Most classification schemes for medical images are focused on specific areas and consider mainly the medical image types (modalities), imaged anatomy, and view, and merge them into a single descriptor or classification hierarchy. Furthermore, they often ignore other important image types such as biological images, statistical figures, flowcharts, and diagrams that frequently occur in the biomedical literature. Most of the current classifications have also been created for radiology images, which are not the only types to be taken into account. With Open Access becoming increasingly widespread particularly in medicine, images from the biomedical literature are more easily available for use. Visual information from these images and knowledge that an image is of a specific type or medical modality could enrich retrieval. This enrichment is hampered by the lack of a commonly agreed image classification scheme. This paper presents a hierarchy for classification of biomedical illustrations with the goal of using it for visual classification and thus as a basis for retrieval. The proposed hierarchy is based on relevant parts of existing terminologies, such as the IRMA-code (Image Retrieval in Medical Applications), ad hoc classifications and hierarchies used in imageCLEF (Image retrieval task at the Cross-Language Evaluation Forum) and NLM's (National Library of Medicine) OpenI. Furtheron, mappings to NLM's MeSH (Medical Subject Headings), RSNA's RadLex (Radiological Society of North America, Radiology Lexicon), and the IRMA code are also attempted for relevant image types. Advantages derived from such hierarchical classification for medical image retrieval are being evaluated through benchmarks such as imageCLEF, and R&D systems such as NLM's OpenI. The goal is to extend this hierarchy progressively and (through adding image types occurring in the biomedical literature) to have a terminology for visual image classification based on image types distinguishable by visual means and occurring in the medical open access literature.

Contour classification in thermographic images for detection of breast cancer

NASA Astrophysics Data System (ADS)

Okuniewski, Rafał; Nowak, Robert M.; Cichosz, Paweł; Jagodziński, Dariusz; Matysiewicz, Mateusz; Neumann, Łukasz; Oleszkiewicz, Witold

2016-09-01

Thermographic images of breast taken by the Braster device are uploaded into web application which uses different classification algorithms to automatically decide whether a patient should be more thoroughly examined. This article presents the approach to the task of classifying contours visible on thermographic images of breast taken by the Braster device in order to make the decision about the existence of cancerous tumors in breast. It presents the results of the researches conducted on the different classification algorithms.

Discriminative Nonlinear Analysis Operator Learning: When Cosparse Model Meets Image Classification.

PubMed

Wen, Zaidao; Hou, Biao; Jiao, Licheng

2017-05-03

Linear synthesis model based dictionary learning framework has achieved remarkable performances in image classification in the last decade. Behaved as a generative feature model, it however suffers from some intrinsic deficiencies. In this paper, we propose a novel parametric nonlinear analysis cosparse model (NACM) with which a unique feature vector will be much more efficiently extracted. Additionally, we derive a deep insight to demonstrate that NACM is capable of simultaneously learning the task adapted feature transformation and regularization to encode our preferences, domain prior knowledge and task oriented supervised information into the features. The proposed NACM is devoted to the classification task as a discriminative feature model and yield a novel discriminative nonlinear analysis operator learning framework (DNAOL). The theoretical analysis and experimental performances clearly demonstrate that DNAOL will not only achieve the better or at least competitive classification accuracies than the state-of-the-art algorithms but it can also dramatically reduce the time complexities in both training and testing phases.

Image Classification Using Biomimetic Pattern Recognition with Convolutional Neural Networks Features

PubMed Central

Huo, Guanying

2017-01-01

As a typical deep-learning model, Convolutional Neural Networks (CNNs) can be exploited to automatically extract features from images using the hierarchical structure inspired by mammalian visual system. For image classification tasks, traditional CNN models employ the softmax function for classification. However, owing to the limited capacity of the softmax function, there are some shortcomings of traditional CNN models in image classification. To deal with this problem, a new method combining Biomimetic Pattern Recognition (BPR) with CNNs is proposed for image classification. BPR performs class recognition by a union of geometrical cover sets in a high-dimensional feature space and therefore can overcome some disadvantages of traditional pattern recognition. The proposed method is evaluated on three famous image classification benchmarks, that is, MNIST, AR, and CIFAR-10. The classification accuracies of the proposed method for the three datasets are 99.01%, 98.40%, and 87.11%, respectively, which are much higher in comparison with the other four methods in most cases. PMID:28316614

Learning semantic histopathological representation for basal cell carcinoma classification

NASA Astrophysics Data System (ADS)

Gutiérrez, Ricardo; Rueda, Andrea; Romero, Eduardo

2013-03-01

Diagnosis of a histopathology glass slide is a complex process that involves accurate recognition of several structures, their function in the tissue and their relation with other structures. The way in which the pathologist represents the image content and the relations between those objects yields a better and accurate diagnoses. Therefore, an appropriate semantic representation of the image content will be useful in several analysis tasks such as cancer classification, tissue retrieval and histopahological image analysis, among others. Nevertheless, to automatically recognize those structures and extract their inner semantic meaning are still very challenging tasks. In this paper we introduce a new semantic representation that allows to describe histopathological concepts suitable for classification. The approach herein identify local concepts using a dictionary learning approach, i.e., the algorithm learns the most representative atoms from a set of random sampled patches, and then models the spatial relations among them by counting the co-occurrence between atoms, while penalizing the spatial distance. The proposed approach was compared with a bag-of-features representation in a tissue classification task. For this purpose, 240 histological microscopical fields of view, 24 per tissue class, were collected. Those images fed a Support Vector Machine classifier per class, using 120 images as train set and the remaining ones for testing, maintaining the same proportion of each concept in the train and test sets. The obtained classification results, averaged from 100 random partitions of training and test sets, shows that our approach is more sensitive in average than the bag-of-features representation in almost 6%.

Multi-level discriminative dictionary learning with application to large scale image classification.

PubMed

Shen, Li; Sun, Gang; Huang, Qingming; Wang, Shuhui; Lin, Zhouchen; Wu, Enhua

2015-10-01

The sparse coding technique has shown flexibility and capability in image representation and analysis. It is a powerful tool in many visual applications. Some recent work has shown that incorporating the properties of task (such as discrimination for classification task) into dictionary learning is effective for improving the accuracy. However, the traditional supervised dictionary learning methods suffer from high computation complexity when dealing with large number of categories, making them less satisfactory in large scale applications. In this paper, we propose a novel multi-level discriminative dictionary learning method and apply it to large scale image classification. Our method takes advantage of hierarchical category correlation to encode multi-level discriminative information. Each internal node of the category hierarchy is associated with a discriminative dictionary and a classification model. The dictionaries at different layers are learnt to capture the information of different scales. Moreover, each node at lower layers also inherits the dictionary of its parent, so that the categories at lower layers can be described with multi-scale information. The learning of dictionaries and associated classification models is jointly conducted by minimizing an overall tree loss. The experimental results on challenging data sets demonstrate that our approach achieves excellent accuracy and competitive computation cost compared with other sparse coding methods for large scale image classification.

Analysis on the application of background parameters on remote sensing classification

NASA Astrophysics Data System (ADS)

Qiao, Y.

Drawing accurate crop cultivation acreage, dynamic monitoring of crops growing and yield forecast are some important applications of remote sensing to agriculture. During the 8th 5-Year Plan period, the task of yield estimation using remote sensing technology for the main crops in major production regions in China once was a subtopic to the national research task titled "Study on Application of Remote sensing Technology". In 21 century in a movement launched by Chinese Ministry of Agriculture to combine high technology to farming production, remote sensing has given full play to farm crops' growth monitoring and yield forecast. And later in 2001 Chinese Ministry of Agriculture entrusted the Northern China Center of Agricultural Remote Sensing to forecast yield of some main crops like wheat, maize and rice in rather short time to supply information for the government decision maker. Present paper is a report for this task. It describes the application of background parameters in image recognition, classification and mapping with focuses on plan of the geo-science's theory, ecological feature and its cartographical objects or scale, the study of phrenology for image optimal time for classification of the ground objects, the analysis of optimal waveband composition and the application of background data base to spatial information recognition ;The research based on the knowledge of background parameters is indispensable for improving the accuracy of image classification and mapping quality and won a secondary reward of tech-science achievement from Chinese Ministry of Agriculture. Keywords: Spatial image; Classification; Background parameter

Exploring Deep Learning and Sparse Matrix Format Selection

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

Zhao, Y.; Liao, C.; Shen, X.

We proposed to explore the use of Deep Neural Networks (DNN) for addressing the longstanding barriers. The recent rapid progress of DNN technology has created a large impact in many fields, which has significantly improved the prediction accuracy over traditional machine learning techniques in image classifications, speech recognitions, machine translations, and so on. To some degree, these tasks resemble the decision makings in many HPC tasks, including the aforementioned format selection for SpMV and linear solver selection. For instance, sparse matrix format selection is akin to image classification—such as, to tell whether an image contains a dog or a cat;more » in both problems, the right decisions are primarily determined by the spatial patterns of the elements in an input. For image classification, the patterns are of pixels, and for sparse matrix format selection, they are of non-zero elements. DNN could be naturally applied if we regard a sparse matrix as an image and the format selection or solver selection as classification problems.« less

Classification of time-series images using deep convolutional neural networks

NASA Astrophysics Data System (ADS)

Hatami, Nima; Gavet, Yann; Debayle, Johan

2018-04-01

Convolutional Neural Networks (CNN) has achieved a great success in image recognition task by automatically learning a hierarchical feature representation from raw data. While the majority of Time-Series Classification (TSC) literature is focused on 1D signals, this paper uses Recurrence Plots (RP) to transform time-series into 2D texture images and then take advantage of the deep CNN classifier. Image representation of time-series introduces different feature types that are not available for 1D signals, and therefore TSC can be treated as texture image recognition task. CNN model also allows learning different levels of representations together with a classifier, jointly and automatically. Therefore, using RP and CNN in a unified framework is expected to boost the recognition rate of TSC. Experimental results on the UCR time-series classification archive demonstrate competitive accuracy of the proposed approach, compared not only to the existing deep architectures, but also to the state-of-the art TSC algorithms.

Optimizing spectral CT parameters for material classification tasks

NASA Astrophysics Data System (ADS)

Rigie, D. S.; La Rivière, P. J.

2016-06-01

In this work, we propose a framework for optimizing spectral CT imaging parameters and hardware design with regard to material classification tasks. Compared with conventional CT, many more parameters must be considered when designing spectral CT systems and protocols. These choices will impact material classification performance in a non-obvious, task-dependent way with direct implications for radiation dose reduction. In light of this, we adapt Hotelling Observer formalisms typically applied to signal detection tasks to the spectral CT, material-classification problem. The result is a rapidly computable metric that makes it possible to sweep out many system configurations, generating parameter optimization curves (POC’s) that can be used to select optimal settings. The proposed model avoids restrictive assumptions about the basis-material decomposition (e.g. linearity) and incorporates signal uncertainty with a stochastic object model. This technique is demonstrated on dual-kVp and photon-counting systems for two different, clinically motivated material classification tasks (kidney stone classification and plaque removal). We show that the POC’s predicted with the proposed analytic model agree well with those derived from computationally intensive numerical simulation studies.

Optimizing Spectral CT Parameters for Material Classification Tasks

PubMed Central

Rigie, D. S.; La Rivière, P. J.

2017-01-01

In this work, we propose a framework for optimizing spectral CT imaging parameters and hardware design with regard to material classification tasks. Compared with conventional CT, many more parameters must be considered when designing spectral CT systems and protocols. These choices will impact material classification performance in a non-obvious, task-dependent way with direct implications for radiation dose reduction. In light of this, we adapt Hotelling Observer formalisms typically applied to signal detection tasks to the spectral CT, material-classification problem. The result is a rapidly computable metric that makes it possible to sweep out many system configurations, generating parameter optimization curves (POC’s) that can be used to select optimal settings. The proposed model avoids restrictive assumptions about the basis-material decomposition (e.g. linearity) and incorporates signal uncertainty with a stochastic object model. This technique is demonstrated on dual-kVp and photon-counting systems for two different, clinically motivated material classification tasks (kidney stone classification and plaque removal). We show that the POC’s predicted with the proposed analytic model agree well with those derived from computationally intensive numerical simulation studies. PMID:27227430

Feature extraction based on extended multi-attribute profiles and sparse autoencoder for remote sensing image classification

NASA Astrophysics Data System (ADS)

Teffahi, Hanane; Yao, Hongxun; Belabid, Nasreddine; Chaib, Souleyman

2018-02-01

The satellite images with very high spatial resolution have been recently widely used in image classification topic as it has become challenging task in remote sensing field. Due to a number of limitations such as the redundancy of features and the high dimensionality of the data, different classification methods have been proposed for remote sensing images classification particularly the methods using feature extraction techniques. This paper propose a simple efficient method exploiting the capability of extended multi-attribute profiles (EMAP) with sparse autoencoder (SAE) for remote sensing image classification. The proposed method is used to classify various remote sensing datasets including hyperspectral and multispectral images by extracting spatial and spectral features based on the combination of EMAP and SAE by linking them to kernel support vector machine (SVM) for classification. Experiments on new hyperspectral image "Huston data" and multispectral image "Washington DC data" shows that this new scheme can achieve better performance of feature learning than the primitive features, traditional classifiers and ordinary autoencoder and has huge potential to achieve higher accuracy for classification in short running time.

Skin Lesion Analysis towards Melanoma Detection Using Deep Learning Network

PubMed Central

2018-01-01

Skin lesions are a severe disease globally. Early detection of melanoma in dermoscopy images significantly increases the survival rate. However, the accurate recognition of melanoma is extremely challenging due to the following reasons: low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc. Hence, reliable automatic detection of skin tumors is very useful to increase the accuracy and efficiency of pathologists. In this paper, we proposed two deep learning methods to address three main tasks emerging in the area of skin lesion image processing, i.e., lesion segmentation (task 1), lesion dermoscopic feature extraction (task 2) and lesion classification (task 3). A deep learning framework consisting of two fully convolutional residual networks (FCRN) is proposed to simultaneously produce the segmentation result and the coarse classification result. A lesion index calculation unit (LICU) is developed to refine the coarse classification results by calculating the distance heat-map. A straight-forward CNN is proposed for the dermoscopic feature extraction task. The proposed deep learning frameworks were evaluated on the ISIC 2017 dataset. Experimental results show the promising accuracies of our frameworks, i.e., 0.753 for task 1, 0.848 for task 2 and 0.912 for task 3 were achieved. PMID:29439500

Optimization of breast mass classification using sequential forward floating selection (SFFS) and a support vector machine (SVM) model

PubMed Central

Tan, Maxine; Pu, Jiantao; Zheng, Bin

2014-01-01

Purpose: Improving radiologists’ performance in classification between malignant and benign breast lesions is important to increase cancer detection sensitivity and reduce false-positive recalls. For this purpose, developing computer-aided diagnosis (CAD) schemes has been attracting research interest in recent years. In this study, we investigated a new feature selection method for the task of breast mass classification. Methods: We initially computed 181 image features based on mass shape, spiculation, contrast, presence of fat or calcifications, texture, isodensity, and other morphological features. From this large image feature pool, we used a sequential forward floating selection (SFFS)-based feature selection method to select relevant features, and analyzed their performance using a support vector machine (SVM) model trained for the classification task. On a database of 600 benign and 600 malignant mass regions of interest (ROIs), we performed the study using a ten-fold cross-validation method. Feature selection and optimization of the SVM parameters were conducted on the training subsets only. Results: The area under the receiver operating characteristic curve (AUC) = 0.805±0.012 was obtained for the classification task. The results also showed that the most frequently-selected features by the SFFS-based algorithm in 10-fold iterations were those related to mass shape, isodensity and presence of fat, which are consistent with the image features frequently used by radiologists in the clinical environment for mass classification. The study also indicated that accurately computing mass spiculation features from the projection mammograms was difficult, and failed to perform well for the mass classification task due to tissue overlap within the benign mass regions. Conclusions: In conclusion, this comprehensive feature analysis study provided new and valuable information for optimizing computerized mass classification schemes that may have potential to be useful as a “second reader” in future clinical practice. PMID:24664267

A Wavelet Polarization Decomposition Net Model for Polarimetric SAR Image Classification

NASA Astrophysics Data System (ADS)

He, Chu; Ou, Dan; Yang, Teng; Wu, Kun; Liao, Mingsheng; Chen, Erxue

2014-11-01

In this paper, a deep model based on wavelet texture has been proposed for Polarimetric Synthetic Aperture Radar (PolSAR) image classification inspired by recent successful deep learning method. Our model is supposed to learn powerful and informative representations to improve the generalization ability for the complex scene classification tasks. Given the influence of speckle noise in Polarimetric SAR image, wavelet polarization decomposition is applied first to obtain basic and discriminative texture features which are then embedded into a Deep Neural Network (DNN) in order to compose multi-layer higher representations. We demonstrate that the model can produce a powerful representation which can capture some untraceable information from Polarimetric SAR images and show a promising achievement in comparison with other traditional SAR image classification methods for the SAR image dataset.

A comparative study for chest radiograph image retrieval using binary texture and deep learning classification.

PubMed

Anavi, Yaron; Kogan, Ilya; Gelbart, Elad; Geva, Ofer; Greenspan, Hayit

2015-08-01

In this work various approaches are investigated for X-ray image retrieval and specifically chest pathology retrieval. Given a query image taken from a data set of 443 images, the objective is to rank images according to similarity. Different features, including binary features, texture features, and deep learning (CNN) features are examined. In addition, two approaches are investigated for the retrieval task. One approach is based on the distance of image descriptors using the above features (hereon termed the "descriptor"-based approach); the second approach ("classification"-based approach) is based on a probability descriptor, generated by a pair-wise classification of each two classes (pathologies) and their decision values using an SVM classifier. Best results are achieved using deep learning features in a classification scheme.

Investigating the Potential of Deep Neural Networks for Large-Scale Classification of Very High Resolution Satellite Images

NASA Astrophysics Data System (ADS)

Postadjian, T.; Le Bris, A.; Sahbi, H.; Mallet, C.

2017-05-01

Semantic classification is a core remote sensing task as it provides the fundamental input for land-cover map generation. The very recent literature has shown the superior performance of deep convolutional neural networks (DCNN) for many classification tasks including the automatic analysis of Very High Spatial Resolution (VHR) geospatial images. Most of the recent initiatives have focused on very high discrimination capacity combined with accurate object boundary retrieval. Therefore, current architectures are perfectly tailored for urban areas over restricted areas but not designed for large-scale purposes. This paper presents an end-to-end automatic processing chain, based on DCNNs, that aims at performing large-scale classification of VHR satellite images (here SPOT 6/7). Since this work assesses, through various experiments, the potential of DCNNs for country-scale VHR land-cover map generation, a simple yet effective architecture is proposed, efficiently discriminating the main classes of interest (namely buildings, roads, water, crops, vegetated areas) by exploiting existing VHR land-cover maps for training.

Computer assisted optical biopsy for colorectal polyps

NASA Astrophysics Data System (ADS)

Navarro-Avila, Fernando J.; Saint-Hill-Febles, Yadira; Renner, Janis; Klare, Peter; von Delius, Stefan; Navab, Nassir; Mateus, Diana

2017-03-01

We propose a method for computer-assisted optical biopsy for colorectal polyps, with the final goal of assisting the medical expert during the colonoscopy. In particular, we target the problem of automatic classification of polyp images in two classes: adenomatous vs non-adenoma. Our approach is based on recent advancements in convolutional neural networks (CNN) for image representation. In the paper, we describe and compare four different methodologies to address the binary classification task: a baseline with classical features and a Random Forest classifier, two methods based on features obtained from a pre-trained network, and finally, the end-to-end training of a CNN. With the pre-trained network, we show the feasibility of transferring a feature extraction mechanism trained on millions of natural images, to the task of classifying adenomatous polyps. We then demonstrate further performance improvements when training the CNN for our specific classification task. In our study, 776 polyp images were acquired and histologically analyzed after polyp resection. We report a performance increase of the CNN-based approaches with respect to both, the conventional engineered features and to a state-of-the-art method based on videos and 3D shape features.

Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases

PubMed Central

Janowczyk, Andrew; Madabhushi, Anant

2016-01-01

Background: Deep learning (DL) is a representation learning approach ideally suited for image analysis challenges in digital pathology (DP). The variety of image analysis tasks in the context of DP includes detection and counting (e.g., mitotic events), segmentation (e.g., nuclei), and tissue classification (e.g., cancerous vs. non-cancerous). Unfortunately, issues with slide preparation, variations in staining and scanning across sites, and vendor platforms, as well as biological variance, such as the presentation of different grades of disease, make these image analysis tasks particularly challenging. Traditional approaches, wherein domain-specific cues are manually identified and developed into task-specific “handcrafted” features, can require extensive tuning to accommodate these variances. However, DL takes a more domain agnostic approach combining both feature discovery and implementation to maximally discriminate between the classes of interest. While DL approaches have performed well in a few DP related image analysis tasks, such as detection and tissue classification, the currently available open source tools and tutorials do not provide guidance on challenges such as (a) selecting appropriate magnification, (b) managing errors in annotations in the training (or learning) dataset, and (c) identifying a suitable training set containing information rich exemplars. These foundational concepts, which are needed to successfully translate the DL paradigm to DP tasks, are non-trivial for (i) DL experts with minimal digital histology experience, and (ii) DP and image processing experts with minimal DL experience, to derive on their own, thus meriting a dedicated tutorial. Aims: This paper investigates these concepts through seven unique DP tasks as use cases to elucidate techniques needed to produce comparable, and in many cases, superior to results from the state-of-the-art hand-crafted feature-based classification approaches. Results: Specifically, in this tutorial on DL for DP image analysis, we show how an open source framework (Caffe), with a singular network architecture, can be used to address: (a) nuclei segmentation (F-score of 0.83 across 12,000 nuclei), (b) epithelium segmentation (F-score of 0.84 across 1735 regions), (c) tubule segmentation (F-score of 0.83 from 795 tubules), (d) lymphocyte detection (F-score of 0.90 across 3064 lymphocytes), (e) mitosis detection (F-score of 0.53 across 550 mitotic events), (f) invasive ductal carcinoma detection (F-score of 0.7648 on 50 k testing patches), and (g) lymphoma classification (classification accuracy of 0.97 across 374 images). Conclusion: This paper represents the largest comprehensive study of DL approaches in DP to date, with over 1200 DP images used during evaluation. The supplemental online material that accompanies this paper consists of step-by-step instructions for the usage of the supplied source code, trained models, and input data. PMID:27563488

Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases.

PubMed

Janowczyk, Andrew; Madabhushi, Anant

2016-01-01

Deep learning (DL) is a representation learning approach ideally suited for image analysis challenges in digital pathology (DP). The variety of image analysis tasks in the context of DP includes detection and counting (e.g., mitotic events), segmentation (e.g., nuclei), and tissue classification (e.g., cancerous vs. non-cancerous). Unfortunately, issues with slide preparation, variations in staining and scanning across sites, and vendor platforms, as well as biological variance, such as the presentation of different grades of disease, make these image analysis tasks particularly challenging. Traditional approaches, wherein domain-specific cues are manually identified and developed into task-specific "handcrafted" features, can require extensive tuning to accommodate these variances. However, DL takes a more domain agnostic approach combining both feature discovery and implementation to maximally discriminate between the classes of interest. While DL approaches have performed well in a few DP related image analysis tasks, such as detection and tissue classification, the currently available open source tools and tutorials do not provide guidance on challenges such as (a) selecting appropriate magnification, (b) managing errors in annotations in the training (or learning) dataset, and (c) identifying a suitable training set containing information rich exemplars. These foundational concepts, which are needed to successfully translate the DL paradigm to DP tasks, are non-trivial for (i) DL experts with minimal digital histology experience, and (ii) DP and image processing experts with minimal DL experience, to derive on their own, thus meriting a dedicated tutorial. This paper investigates these concepts through seven unique DP tasks as use cases to elucidate techniques needed to produce comparable, and in many cases, superior to results from the state-of-the-art hand-crafted feature-based classification approaches. Specifically, in this tutorial on DL for DP image analysis, we show how an open source framework (Caffe), with a singular network architecture, can be used to address: (a) nuclei segmentation (F-score of 0.83 across 12,000 nuclei), (b) epithelium segmentation (F-score of 0.84 across 1735 regions), (c) tubule segmentation (F-score of 0.83 from 795 tubules), (d) lymphocyte detection (F-score of 0.90 across 3064 lymphocytes), (e) mitosis detection (F-score of 0.53 across 550 mitotic events), (f) invasive ductal carcinoma detection (F-score of 0.7648 on 50 k testing patches), and (g) lymphoma classification (classification accuracy of 0.97 across 374 images). This paper represents the largest comprehensive study of DL approaches in DP to date, with over 1200 DP images used during evaluation. The supplemental online material that accompanies this paper consists of step-by-step instructions for the usage of the supplied source code, trained models, and input data.

Attribute-based classification for zero-shot visual object categorization.

PubMed

Lampert, Christoph H; Nickisch, Hannes; Harmeling, Stefan

2014-03-01

We study the problem of object recognition for categories for which we have no training examples, a task also called zero--data or zero-shot learning. This situation has hardly been studied in computer vision research, even though it occurs frequently; the world contains tens of thousands of different object classes, and image collections have been formed and suitably annotated for only a few of them. To tackle the problem, we introduce attribute-based classification: Objects are identified based on a high-level description that is phrased in terms of semantic attributes, such as the object's color or shape. Because the identification of each such property transcends the specific learning task at hand, the attribute classifiers can be prelearned independently, for example, from existing image data sets unrelated to the current task. Afterward, new classes can be detected based on their attribute representation, without the need for a new training phase. In this paper, we also introduce a new data set, Animals with Attributes, of over 30,000 images of 50 animal classes, annotated with 85 semantic attributes. Extensive experiments on this and two more data sets show that attribute-based classification indeed is able to categorize images without access to any training images of the target classes.

Large-area settlement pattern recognition from Landsat-8 data

NASA Astrophysics Data System (ADS)

Wieland, Marc; Pittore, Massimiliano

2016-09-01

The study presents an image processing and analysis pipeline that combines object-based image analysis with a Support Vector Machine to derive a multi-layered settlement product from Landsat-8 data over large areas. 43 image scenes are processed over large parts of Central Asia (Southern Kazakhstan, Kyrgyzstan, Tajikistan and Eastern Uzbekistan). The main tasks tackled by this work include built-up area identification, settlement type classification and urban structure types pattern recognition. Besides commonly used accuracy assessments of the resulting map products, thorough performance evaluations are carried out under varying conditions to tune algorithm parameters and assess their applicability for the given tasks. As part of this, several research questions are being addressed. In particular the influence of the improved spatial and spectral resolution of Landsat-8 on the SVM performance to identify built-up areas and urban structure types are evaluated. Also the influence of an extended feature space including digital elevation model features is tested for mountainous regions. Moreover, the spatial distribution of classification uncertainties is analyzed and compared to the heterogeneity of the building stock within the computational unit of the segments. The study concludes that the information content of Landsat-8 images is sufficient for the tested classification tasks and even detailed urban structures could be extracted with satisfying accuracy. Freely available ancillary settlement point location data could further improve the built-up area classification. Digital elevation features and pan-sharpening could, however, not significantly improve the classification results. The study highlights the importance of dynamically tuned classifier parameters, and underlines the use of Shannon entropy computed from the soft answers of the SVM as a valid measure of the spatial distribution of classification uncertainties.

Application of Sensor Fusion to Improve Uav Image Classification

NASA Astrophysics Data System (ADS)

Jabari, S.; Fathollahi, F.; Zhang, Y.

2017-08-01

Image classification is one of the most important tasks of remote sensing projects including the ones that are based on using UAV images. Improving the quality of UAV images directly affects the classification results and can save a huge amount of time and effort in this area. In this study, we show that sensor fusion can improve image quality which results in increasing the accuracy of image classification. Here, we tested two sensor fusion configurations by using a Panchromatic (Pan) camera along with either a colour camera or a four-band multi-spectral (MS) camera. We use the Pan camera to benefit from its higher sensitivity and the colour or MS camera to benefit from its spectral properties. The resulting images are then compared to the ones acquired by a high resolution single Bayer-pattern colour camera (here referred to as HRC). We assessed the quality of the output images by performing image classification tests. The outputs prove that the proposed sensor fusion configurations can achieve higher accuracies compared to the images of the single Bayer-pattern colour camera. Therefore, incorporating a Pan camera on-board in the UAV missions and performing image fusion can help achieving higher quality images and accordingly higher accuracy classification results.

Evaluating performance of biomedical image retrieval systems – an overview of the medical image retrieval task at ImageCLEF 2004–2013

PubMed Central

Kalpathy-Cramer, Jayashree; de Herrera, Alba García Seco; Demner-Fushman, Dina; Antani, Sameer; Bedrick, Steven; Müller, Henning

2014-01-01

Medical image retrieval and classification have been extremely active research topics over the past 15 years. With the ImageCLEF benchmark in medical image retrieval and classification a standard test bed was created that allows researchers to compare their approaches and ideas on increasingly large and varied data sets including generated ground truth. This article describes the lessons learned in ten evaluations campaigns. A detailed analysis of the data also highlights the value of the resources created. PMID:24746250

A Generic Deep-Learning-Based Approach for Automated Surface Inspection.

PubMed

Ren, Ruoxu; Hung, Terence; Tan, Kay Chen

2018-03-01

Automated surface inspection (ASI) is a challenging task in industry, as collecting training dataset is usually costly and related methods are highly dataset-dependent. In this paper, a generic approach that requires small training data for ASI is proposed. First, this approach builds classifier on the features of image patches, where the features are transferred from a pretrained deep learning network. Next, pixel-wise prediction is obtained by convolving the trained classifier over input image. An experiment on three public and one industrial data set is carried out. The experiment involves two tasks: 1) image classification and 2) defect segmentation. The results of proposed algorithm are compared against several best benchmarks in literature. In the classification tasks, the proposed method improves accuracy by 0.66%-25.50%. In the segmentation tasks, the proposed method reduces error escape rates by 6.00%-19.00% in three defect types and improves accuracies by 2.29%-9.86% in all seven defect types. In addition, the proposed method achieves 0.0% error escape rate in the segmentation task of industrial data.

Image-based deep learning for classification of noise transients in gravitational wave detectors

NASA Astrophysics Data System (ADS)

Razzano, Massimiliano; Cuoco, Elena

2018-05-01

The detection of gravitational waves has inaugurated the era of gravitational astronomy and opened new avenues for the multimessenger study of cosmic sources. Thanks to their sensitivity, the Advanced LIGO and Advanced Virgo interferometers will probe a much larger volume of space and expand the capability of discovering new gravitational wave emitters. The characterization of these detectors is a primary task in order to recognize the main sources of noise and optimize the sensitivity of interferometers. Glitches are transient noise events that can impact the data quality of the interferometers and their classification is an important task for detector characterization. Deep learning techniques are a promising tool for the recognition and classification of glitches. We present a classification pipeline that exploits convolutional neural networks to classify glitches starting from their time-frequency evolution represented as images. We evaluated the classification accuracy on simulated glitches, showing that the proposed algorithm can automatically classify glitches on very fast timescales and with high accuracy, thus providing a promising tool for online detector characterization.

Multitask SVM learning for remote sensing data classification

NASA Astrophysics Data System (ADS)

Leiva-Murillo, Jose M.; Gómez-Chova, Luis; Camps-Valls, Gustavo

2010-10-01

Many remote sensing data processing problems are inherently constituted by several tasks that can be solved either individually or jointly. For instance, each image in a multitemporal classification setting could be taken as an individual task but relation to previous acquisitions should be properly considered. In such problems, different modalities of the data (temporal, spatial, angular) gives rise to changes between the training and test distributions, which constitutes a difficult learning problem known as covariate shift. Multitask learning methods aim at jointly solving a set of prediction problems in an efficient way by sharing information across tasks. This paper presents a novel kernel method for multitask learning in remote sensing data classification. The proposed method alleviates the dataset shift problem by imposing cross-information in the classifiers through matrix regularization. We consider the support vector machine (SVM) as core learner and two regularization schemes are introduced: 1) the Euclidean distance of the predictors in the Hilbert space; and 2) the inclusion of relational operators between tasks. Experiments are conducted in the challenging remote sensing problems of cloud screening from multispectral MERIS images and for landmine detection.

Deep multi-scale convolutional neural network for hyperspectral image classification

NASA Astrophysics Data System (ADS)

Zhang, Feng-zhe; Yang, Xia

2018-04-01

In this paper, we proposed a multi-scale convolutional neural network for hyperspectral image classification task. Firstly, compared with conventional convolution, we utilize multi-scale convolutions, which possess larger respective fields, to extract spectral features of hyperspectral image. We design a deep neural network with a multi-scale convolution layer which contains 3 different convolution kernel sizes. Secondly, to avoid overfitting of deep neural network, dropout is utilized, which randomly sleeps neurons, contributing to improve the classification accuracy a bit. In addition, new skills like ReLU in deep learning is utilized in this paper. We conduct experiments on University of Pavia and Salinas datasets, and obtained better classification accuracy compared with other methods.

Deep Convolutional Neural Network-Based Early Automated Detection of Diabetic Retinopathy Using Fundus Image.

PubMed

Xu, Kele; Feng, Dawei; Mi, Haibo

2017-11-23

The automatic detection of diabetic retinopathy is of vital importance, as it is the main cause of irreversible vision loss in the working-age population in the developed world. The early detection of diabetic retinopathy occurrence can be very helpful for clinical treatment; although several different feature extraction approaches have been proposed, the classification task for retinal images is still tedious even for those trained clinicians. Recently, deep convolutional neural networks have manifested superior performance in image classification compared to previous handcrafted feature-based image classification methods. Thus, in this paper, we explored the use of deep convolutional neural network methodology for the automatic classification of diabetic retinopathy using color fundus image, and obtained an accuracy of 94.5% on our dataset, outperforming the results obtained by using classical approaches.

Parallel processing considerations for image recognition tasks

NASA Astrophysics Data System (ADS)

Simske, Steven J.

2011-01-01

Many image recognition tasks are well-suited to parallel processing. The most obvious example is that many imaging tasks require the analysis of multiple images. From this standpoint, then, parallel processing need be no more complicated than assigning individual images to individual processors. However, there are three less trivial categories of parallel processing that will be considered in this paper: parallel processing (1) by task; (2) by image region; and (3) by meta-algorithm. Parallel processing by task allows the assignment of multiple workflows-as diverse as optical character recognition [OCR], document classification and barcode reading-to parallel pipelines. This can substantially decrease time to completion for the document tasks. For this approach, each parallel pipeline is generally performing a different task. Parallel processing by image region allows a larger imaging task to be sub-divided into a set of parallel pipelines, each performing the same task but on a different data set. This type of image analysis is readily addressed by a map-reduce approach. Examples include document skew detection and multiple face detection and tracking. Finally, parallel processing by meta-algorithm allows different algorithms to be deployed on the same image simultaneously. This approach may result in improved accuracy.

Recurrent neural networks for breast lesion classification based on DCE-MRIs

NASA Astrophysics Data System (ADS)

Antropova, Natasha; Huynh, Benjamin; Giger, Maryellen

2018-02-01

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) plays a significant role in breast cancer screening, cancer staging, and monitoring response to therapy. Recently, deep learning methods are being rapidly incorporated in image-based breast cancer diagnosis and prognosis. However, most of the current deep learning methods make clinical decisions based on 2-dimentional (2D) or 3D images and are not well suited for temporal image data. In this study, we develop a deep learning methodology that enables integration of clinically valuable temporal components of DCE-MRIs into deep learning-based lesion classification. Our work is performed on a database of 703 DCE-MRI cases for the task of distinguishing benign and malignant lesions, and uses the area under the ROC curve (AUC) as the performance metric in conducting that task. We train a recurrent neural network, specifically a long short-term memory network (LSTM), on sequences of image features extracted from the dynamic MRI sequences. These features are extracted with VGGNet, a convolutional neural network pre-trained on a large dataset of natural images ImageNet. The features are obtained from various levels of the network, to capture low-, mid-, and high-level information about the lesion. Compared to a classification method that takes as input only images at a single time-point (yielding an AUC = 0.81 (se = 0.04)), our LSTM method improves lesion classification with an AUC of 0.85 (se = 0.03).

Information based universal feature extraction

NASA Astrophysics Data System (ADS)

Amiri, Mohammad; Brause, Rüdiger

2015-02-01

In many real world image based pattern recognition tasks, the extraction and usage of task-relevant features are the most crucial part of the diagnosis. In the standard approach, they mostly remain task-specific, although humans who perform such a task always use the same image features, trained in early childhood. It seems that universal feature sets exist, but they are not yet systematically found. In our contribution, we tried to find those universal image feature sets that are valuable for most image related tasks. In our approach, we trained a neural network by natural and non-natural images of objects and background, using a Shannon information-based algorithm and learning constraints. The goal was to extract those features that give the most valuable information for classification of visual objects hand-written digits. This will give a good start and performance increase for all other image learning tasks, implementing a transfer learning approach. As result, in our case we found that we could indeed extract features which are valid in all three kinds of tasks.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification

NASA Astrophysics Data System (ADS)

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-12-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification.

PubMed

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-12-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value.

A Parallel Adaboost-Backpropagation Neural Network for Massive Image Dataset Classification

PubMed Central

Cao, Jianfang; Chen, Lichao; Wang, Min; Shi, Hao; Tian, Yun

2016-01-01

Image classification uses computers to simulate human understanding and cognition of images by automatically categorizing images. This study proposes a faster image classification approach that parallelizes the traditional Adaboost-Backpropagation (BP) neural network using the MapReduce parallel programming model. First, we construct a strong classifier by assembling the outputs of 15 BP neural networks (which are individually regarded as weak classifiers) based on the Adaboost algorithm. Second, we design Map and Reduce tasks for both the parallel Adaboost-BP neural network and the feature extraction algorithm. Finally, we establish an automated classification model by building a Hadoop cluster. We use the Pascal VOC2007 and Caltech256 datasets to train and test the classification model. The results are superior to those obtained using traditional Adaboost-BP neural network or parallel BP neural network approaches. Our approach increased the average classification accuracy rate by approximately 14.5% and 26.0% compared to the traditional Adaboost-BP neural network and parallel BP neural network, respectively. Furthermore, the proposed approach requires less computation time and scales very well as evaluated by speedup, sizeup and scaleup. The proposed approach may provide a foundation for automated large-scale image classification and demonstrates practical value. PMID:27905520

Multilevel image recognition using discriminative patches and kernel covariance descriptor

NASA Astrophysics Data System (ADS)

Lu, Le; Yao, Jianhua; Turkbey, Evrim; Summers, Ronald M.

2014-03-01

Computer-aided diagnosis of medical images has emerged as an important tool to objectively improve the performance, accuracy and consistency for clinical workflow. To computerize the medical image diagnostic recognition problem, there are three fundamental problems: where to look (i.e., where is the region of interest from the whole image/volume), image feature description/encoding, and similarity metrics for classification or matching. In this paper, we exploit the motivation, implementation and performance evaluation of task-driven iterative, discriminative image patch mining; covariance matrix based descriptor via intensity, gradient and spatial layout; and log-Euclidean distance kernel for support vector machine, to address these three aspects respectively. To cope with often visually ambiguous image patterns for the region of interest in medical diagnosis, discovery of multilabel selective discriminative patches is desired. Covariance of several image statistics summarizes their second order interactions within an image patch and is proved as an effective image descriptor, with low dimensionality compared with joint statistics and fast computation regardless of the patch size. We extensively evaluate two extended Gaussian kernels using affine-invariant Riemannian metric or log-Euclidean metric with support vector machines (SVM), on two medical image classification problems of degenerative disc disease (DDD) detection on cortical shell unwrapped CT maps and colitis detection on CT key images. The proposed approach is validated with promising quantitative results on these challenging tasks. Our experimental findings and discussion also unveil some interesting insights on the covariance feature composition with or without spatial layout for classification and retrieval, and different kernel constructions for SVM. This will also shed some light on future work using covariance feature and kernel classification for medical image analysis.

Minimization of annotation work: diagnosis of mammographic masses via active learning

NASA Astrophysics Data System (ADS)

Zhao, Yu; Zhang, Jingyang; Xie, Hongzhi; Zhang, Shuyang; Gu, Lixu

2018-06-01

The prerequisite for establishing an effective prediction system for mammographic diagnosis is the annotation of each mammographic image. The manual annotation work is time-consuming and laborious, which becomes a great hindrance for researchers. In this article, we propose a novel active learning algorithm that can adequately address this problem, leading to the minimization of the labeling costs on the premise of guaranteed performance. Our proposed method is different from the existing active learning methods designed for the general problem as it is specifically designed for mammographic images. Through its modified discriminant functions and improved sample query criteria, the proposed method can fully utilize the pairing of mammographic images and select the most valuable images from both the mediolateral and craniocaudal views. Moreover, in order to extend active learning to the ordinal regression problem, which has no precedent in existing studies, but is essential for mammographic diagnosis (mammographic diagnosis is not only a classification task, but also an ordinal regression task for predicting an ordinal variable, viz. the malignancy risk of lesions), multiple sample query criteria need to be taken into consideration simultaneously. We formulate it as a criteria integration problem and further present an algorithm based on self-adaptive weighted rank aggregation to achieve a good solution. The efficacy of the proposed method was demonstrated on thousands of mammographic images from the digital database for screening mammography. The labeling costs of obtaining optimal performance in the classification and ordinal regression task respectively fell to 33.8 and 19.8 percent of their original costs. The proposed method also generated 1228 wins, 369 ties and 47 losses for the classification task, and 1933 wins, 258 ties and 185 losses for the ordinal regression task compared to the other state-of-the-art active learning algorithms. By taking the particularities of mammographic images, the proposed AL method can indeed reduce the manual annotation work to a great extent without sacrificing the performance of the prediction system for mammographic diagnosis.

Minimization of annotation work: diagnosis of mammographic masses via active learning.

PubMed

Zhao, Yu; Zhang, Jingyang; Xie, Hongzhi; Zhang, Shuyang; Gu, Lixu

2018-05-22

The prerequisite for establishing an effective prediction system for mammographic diagnosis is the annotation of each mammographic image. The manual annotation work is time-consuming and laborious, which becomes a great hindrance for researchers. In this article, we propose a novel active learning algorithm that can adequately address this problem, leading to the minimization of the labeling costs on the premise of guaranteed performance. Our proposed method is different from the existing active learning methods designed for the general problem as it is specifically designed for mammographic images. Through its modified discriminant functions and improved sample query criteria, the proposed method can fully utilize the pairing of mammographic images and select the most valuable images from both the mediolateral and craniocaudal views. Moreover, in order to extend active learning to the ordinal regression problem, which has no precedent in existing studies, but is essential for mammographic diagnosis (mammographic diagnosis is not only a classification task, but also an ordinal regression task for predicting an ordinal variable, viz. the malignancy risk of lesions), multiple sample query criteria need to be taken into consideration simultaneously. We formulate it as a criteria integration problem and further present an algorithm based on self-adaptive weighted rank aggregation to achieve a good solution. The efficacy of the proposed method was demonstrated on thousands of mammographic images from the digital database for screening mammography. The labeling costs of obtaining optimal performance in the classification and ordinal regression task respectively fell to 33.8 and 19.8 percent of their original costs. The proposed method also generated 1228 wins, 369 ties and 47 losses for the classification task, and 1933 wins, 258 ties and 185 losses for the ordinal regression task compared to the other state-of-the-art active learning algorithms. By taking the particularities of mammographic images, the proposed AL method can indeed reduce the manual annotation work to a great extent without sacrificing the performance of the prediction system for mammographic diagnosis.

Singular spectrum decomposition of Bouligand-Minkowski fractal descriptors: an application to the classification of texture Images

NASA Astrophysics Data System (ADS)

Florindo, João. Batista

2018-04-01

This work proposes the use of Singular Spectrum Analysis (SSA) for the classification of texture images, more specifically, to enhance the performance of the Bouligand-Minkowski fractal descriptors in this task. Fractal descriptors are known to be a powerful approach to model and particularly identify complex patterns in natural images. Nevertheless, the multiscale analysis involved in those descriptors makes them highly correlated. Although other attempts to address this point was proposed in the literature, none of them investigated the relation between the fractal correlation and the well-established analysis employed in time series. And SSA is one of the most powerful techniques for this purpose. The proposed method was employed for the classification of benchmark texture images and the results were compared with other state-of-the-art classifiers, confirming the potential of this analysis in image classification.

Bag-of-visual-ngrams for histopathology image classification

NASA Astrophysics Data System (ADS)

López-Monroy, A. Pastor; Montes-y-Gómez, Manuel; Escalante, Hugo Jair; Cruz-Roa, Angel; González, Fabio A.

2013-11-01

This paper describes an extension of the Bag-of-Visual-Words (BoVW) representation for image categorization (IC) of histophatology images. This representation is one of the most used approaches in several high-level computer vision tasks. However, the BoVW representation has an important limitation: the disregarding of spatial information among visual words. This information may be useful to capture discriminative visual-patterns in specific computer vision tasks. In order to overcome this problem we propose the use of visual n-grams. N-grams based-representations are very popular in the field of natural language processing (NLP), in particular within text mining and information retrieval. We propose building a codebook of n-grams and then representing images by histograms of visual n-grams. We evaluate our proposal in the challenging task of classifying histopathology images. The novelty of our proposal lies in the fact that we use n-grams as attributes for a classification model (together with visual-words, i.e., 1-grams). This is common practice within NLP, although, to the best of our knowledge, this idea has not been explored yet within computer vision. We report experimental results in a database of histopathology images where our proposed method outperforms the traditional BoVWs formulation.

Collaborative classification of hyperspectral and visible images with convolutional neural network

NASA Astrophysics Data System (ADS)

Zhang, Mengmeng; Li, Wei; Du, Qian

2017-10-01

Recent advances in remote sensing technology have made multisensor data available for the same area, and it is well-known that remote sensing data processing and analysis often benefit from multisource data fusion. Specifically, low spatial resolution of hyperspectral imagery (HSI) degrades the quality of the subsequent classification task while using visible (VIS) images with high spatial resolution enables high-fidelity spatial analysis. A collaborative classification framework is proposed to fuse HSI and VIS images for finer classification. First, the convolutional neural network model is employed to extract deep spectral features for HSI classification. Second, effective binarized statistical image features are learned as contextual basis vectors for the high-resolution VIS image, followed by a classifier. The proposed approach employs diversified data in a decision fusion, leading to an integration of the rich spectral information, spatial information, and statistical representation information. In particular, the proposed approach eliminates the potential problems of the curse of dimensionality and excessive computation time. The experiments evaluated on two standard data sets demonstrate better classification performance offered by this framework.

Exploiting ensemble learning for automatic cataract detection and grading.

PubMed

Yang, Ji-Jiang; Li, Jianqiang; Shen, Ruifang; Zeng, Yang; He, Jian; Bi, Jing; Li, Yong; Zhang, Qinyan; Peng, Lihui; Wang, Qing

2016-02-01

Cataract is defined as a lenticular opacity presenting usually with poor visual acuity. It is one of the most common causes of visual impairment worldwide. Early diagnosis demands the expertise of trained healthcare professionals, which may present a barrier to early intervention due to underlying costs. To date, studies reported in the literature utilize a single learning model for retinal image classification in grading cataract severity. We present an ensemble learning based approach as a means to improving diagnostic accuracy. Three independent feature sets, i.e., wavelet-, sketch-, and texture-based features, are extracted from each fundus image. For each feature set, two base learning models, i.e., Support Vector Machine and Back Propagation Neural Network, are built. Then, the ensemble methods, majority voting and stacking, are investigated to combine the multiple base learning models for final fundus image classification. Empirical experiments are conducted for cataract detection (two-class task, i.e., cataract or non-cataractous) and cataract grading (four-class task, i.e., non-cataractous, mild, moderate or severe) tasks. The best performance of the ensemble classifier is 93.2% and 84.5% in terms of the correct classification rates for cataract detection and grading tasks, respectively. The results demonstrate that the ensemble classifier outperforms the single learning model significantly, which also illustrates the effectiveness of the proposed approach. Copyright © 2015 Elsevier Ireland Ltd. All rights reserved.

Pet fur color and texture classification

NASA Astrophysics Data System (ADS)

Yen, Jonathan; Mukherjee, Debarghar; Lim, SukHwan; Tretter, Daniel

2007-01-01

Object segmentation is important in image analysis for imaging tasks such as image rendering and image retrieval. Pet owners have been known to be quite vocal about how important it is to render their pets perfectly. We present here an algorithm for pet (mammal) fur color classification and an algorithm for pet (animal) fur texture classification. Per fur color classification can be applied as a necessary condition for identifying the regions in an image that may contain pets much like the skin tone classification for human flesh detection. As a result of the evolution, fur coloration of all mammals is caused by a natural organic pigment called Melanin and Melanin has only very limited color ranges. We have conducted a statistical analysis and concluded that mammal fur colors can be only in levels of gray or in two colors after the proper color quantization. This pet fur color classification algorithm has been applied for peteye detection. We also present here an algorithm for animal fur texture classification using the recently developed multi-resolution directional sub-band Contourlet transform. The experimental results are very promising as these transforms can identify regions of an image that may contain fur of mammals, scale of reptiles and feather of birds, etc. Combining the color and texture classification, one can have a set of strong classifiers for identifying possible animals in an image.

Ethnicity identification from face images

NASA Astrophysics Data System (ADS)

Lu, Xiaoguang; Jain, Anil K.

2004-08-01

Human facial images provide the demographic information, such as ethnicity and gender. Conversely, ethnicity and gender also play an important role in face-related applications. Image-based ethnicity identification problem is addressed in a machine learning framework. The Linear Discriminant Analysis (LDA) based scheme is presented for the two-class (Asian vs. non-Asian) ethnicity classification task. Multiscale analysis is applied to the input facial images. An ensemble framework, which integrates the LDA analysis for the input face images at different scales, is proposed to further improve the classification performance. The product rule is used as the combination strategy in the ensemble. Experimental results based on a face database containing 263 subjects (2,630 face images, with equal balance between the two classes) are promising, indicating that LDA and the proposed ensemble framework have sufficient discriminative power for the ethnicity classification problem. The normalized ethnicity classification scores can be helpful in the facial identity recognition. Useful as a "soft" biometric, face matching scores can be updated based on the output of ethnicity classification module. In other words, ethnicity classifier does not have to be perfect to be useful in practice.

Decoding memory features from hippocampal spiking activities using sparse classification models.

PubMed

Dong Song; Hampson, Robert E; Robinson, Brian S; Marmarelis, Vasilis Z; Deadwyler, Sam A; Berger, Theodore W

2016-08-01

To understand how memory information is encoded in the hippocampus, we build classification models to decode memory features from hippocampal CA3 and CA1 spatio-temporal patterns of spikes recorded from epilepsy patients performing a memory-dependent delayed match-to-sample task. The classification model consists of a set of B-spline basis functions for extracting memory features from the spike patterns, and a sparse logistic regression classifier for generating binary categorical output of memory features. Results show that classification models can extract significant amount of memory information with respects to types of memory tasks and categories of sample images used in the task, despite the high level of variability in prediction accuracy due to the small sample size. These results support the hypothesis that memories are encoded in the hippocampal activities and have important implication to the development of hippocampal memory prostheses.

Visual affective classification by combining visual and text features.

PubMed

Liu, Ningning; Wang, Kai; Jin, Xin; Gao, Boyang; Dellandréa, Emmanuel; Chen, Liming

2017-01-01

Affective analysis of images in social networks has drawn much attention, and the texts surrounding images are proven to provide valuable semantic meanings about image content, which can hardly be represented by low-level visual features. In this paper, we propose a novel approach for visual affective classification (VAC) task. This approach combines visual representations along with novel text features through a fusion scheme based on Dempster-Shafer (D-S) Evidence Theory. Specifically, we not only investigate different types of visual features and fusion methods for VAC, but also propose textual features to effectively capture emotional semantics from the short text associated to images based on word similarity. Experiments are conducted on three public available databases: the International Affective Picture System (IAPS), the Artistic Photos and the MirFlickr Affect set. The results demonstrate that the proposed approach combining visual and textual features provides promising results for VAC task.

Visual affective classification by combining visual and text features

PubMed Central

Liu, Ningning; Wang, Kai; Jin, Xin; Gao, Boyang; Dellandréa, Emmanuel; Chen, Liming

2017-01-01

Affective analysis of images in social networks has drawn much attention, and the texts surrounding images are proven to provide valuable semantic meanings about image content, which can hardly be represented by low-level visual features. In this paper, we propose a novel approach for visual affective classification (VAC) task. This approach combines visual representations along with novel text features through a fusion scheme based on Dempster-Shafer (D-S) Evidence Theory. Specifically, we not only investigate different types of visual features and fusion methods for VAC, but also propose textual features to effectively capture emotional semantics from the short text associated to images based on word similarity. Experiments are conducted on three public available databases: the International Affective Picture System (IAPS), the Artistic Photos and the MirFlickr Affect set. The results demonstrate that the proposed approach combining visual and textual features provides promising results for VAC task. PMID:28850566

Image aesthetic quality evaluation using convolution neural network embedded learning

NASA Astrophysics Data System (ADS)

Li, Yu-xin; Pu, Yuan-yuan; Xu, Dan; Qian, Wen-hua; Wang, Li-peng

2017-11-01

A way of embedded learning convolution neural network (ELCNN) based on the image content is proposed to evaluate the image aesthetic quality in this paper. Our approach can not only solve the problem of small-scale data but also score the image aesthetic quality. First, we chose Alexnet and VGG_S to compare for confirming which is more suitable for this image aesthetic quality evaluation task. Second, to further boost the image aesthetic quality classification performance, we employ the image content to train aesthetic quality classification models. But the training samples become smaller and only using once fine-tuning cannot make full use of the small-scale data set. Third, to solve the problem in second step, a way of using twice fine-tuning continually based on the aesthetic quality label and content label respective is proposed, the classification probability of the trained CNN models is used to evaluate the image aesthetic quality. The experiments are carried on the small-scale data set of Photo Quality. The experiment results show that the classification accuracy rates of our approach are higher than the existing image aesthetic quality evaluation approaches.

Single-trial EEG RSVP classification using convolutional neural networks

NASA Astrophysics Data System (ADS)

Shamwell, Jared; Lee, Hyungtae; Kwon, Heesung; Marathe, Amar R.; Lawhern, Vernon; Nothwang, William

2016-05-01

Traditionally, Brain-Computer Interfaces (BCI) have been explored as a means to return function to paralyzed or otherwise debilitated individuals. An emerging use for BCIs is in human-autonomy sensor fusion where physiological data from healthy subjects is combined with machine-generated information to enhance the capabilities of artificial systems. While human-autonomy fusion of physiological data and computer vision have been shown to improve classification during visual search tasks, to date these approaches have relied on separately trained classification models for each modality. We aim to improve human-autonomy classification performance by developing a single framework that builds codependent models of human electroencephalograph (EEG) and image data to generate fused target estimates. As a first step, we developed a novel convolutional neural network (CNN) architecture and applied it to EEG recordings of subjects classifying target and non-target image presentations during a rapid serial visual presentation (RSVP) image triage task. The low signal-to-noise ratio (SNR) of EEG inherently limits the accuracy of single-trial classification and when combined with the high dimensionality of EEG recordings, extremely large training sets are needed to prevent overfitting and achieve accurate classification from raw EEG data. This paper explores a new deep CNN architecture for generalized multi-class, single-trial EEG classification across subjects. We compare classification performance from the generalized CNN architecture trained across all subjects to the individualized XDAWN, HDCA, and CSP neural classifiers which are trained and tested on single subjects. Preliminary results show that our CNN meets and slightly exceeds the performance of the other classifiers despite being trained across subjects.

An improved arteriovenous classification method for the early diagnostics of various diseases in retinal image.

PubMed

Xu, Xiayu; Ding, Wenxiang; Abràmoff, Michael D; Cao, Ruofan

2017-04-01

Retinal artery and vein classification is an important task for the automatic computer-aided diagnosis of various eye diseases and systemic diseases. This paper presents an improved supervised artery and vein classification method in retinal image. Intra-image regularization and inter-subject normalization is applied to reduce the differences in feature space. Novel features, including first-order and second-order texture features, are utilized to capture the discriminating characteristics of arteries and veins. The proposed method was tested on the DRIVE dataset and achieved an overall accuracy of 0.923. This retinal artery and vein classification algorithm serves as a potentially important tool for the early diagnosis of various diseases, including diabetic retinopathy and cardiovascular diseases. Copyright © 2017 Elsevier B.V. All rights reserved.

Evolutionary image simplification for lung nodule classification with convolutional neural networks.

PubMed

Lückehe, Daniel; von Voigt, Gabriele

2018-05-29

Understanding decisions of deep learning techniques is important. Especially in the medical field, the reasons for a decision in a classification task are as crucial as the pure classification results. In this article, we propose a new approach to compute relevant parts of a medical image. Knowing the relevant parts makes it easier to understand decisions. In our approach, a convolutional neural network is employed to learn structures of images of lung nodules. Then, an evolutionary algorithm is applied to compute a simplified version of an unknown image based on the learned structures by the convolutional neural network. In the simplified version, irrelevant parts are removed from the original image. In the results, we show simplified images which allow the observer to focus on the relevant parts. In these images, more than 50% of the pixels are simplified. The simplified pixels do not change the meaning of the images based on the learned structures by the convolutional neural network. An experimental analysis shows the potential of the approach. Besides the examples of simplified images, we analyze the run time development. Simplified images make it easier to focus on relevant parts and to find reasons for a decision. The combination of an evolutionary algorithm employing a learned convolutional neural network is well suited for the simplification task. From a research perspective, it is interesting which areas of the images are simplified and which parts are taken as relevant.

Learning discriminative features from RGB-D images for gender and ethnicity identification

NASA Astrophysics Data System (ADS)

Azzakhnini, Safaa; Ballihi, Lahoucine; Aboutajdine, Driss

2016-11-01

The development of sophisticated sensor technologies gave rise to an interesting variety of data. With the appearance of affordable devices, such as the Microsoft Kinect, depth-maps and three-dimensional data became easily accessible. This attracted many computer vision researchers seeking to exploit this information in classification and recognition tasks. In this work, the problem of face classification in the context of RGB images and depth information (RGB-D images) is addressed. The purpose of this paper is to study and compare some popular techniques for gender recognition and ethnicity classification to understand how much depth data can improve the quality of recognition. Furthermore, we investigate which combination of face descriptors, feature selection methods, and learning techniques is best suited to better exploit RGB-D images. The experimental results show that depth data improve the recognition accuracy for gender and ethnicity classification applications in many use cases.

Land-use Scene Classification in High-Resolution Remote Sensing Images by Multiscale Deeply Described Correlatons

NASA Astrophysics Data System (ADS)

Qi, K.; Qingfeng, G.

2017-12-01

With the popular use of High-Resolution Satellite (HRS) images, more and more research efforts have been placed on land-use scene classification. However, it makes the task difficult with HRS images for the complex background and multiple land-cover classes or objects. This article presents a multiscale deeply described correlaton model for land-use scene classification. Specifically, the convolutional neural network is introduced to learn and characterize the local features at different scales. Then, learnt multiscale deep features are explored to generate visual words. The spatial arrangement of visual words is achieved through the introduction of adaptive vector quantized correlograms at different scales. Experiments on two publicly available land-use scene datasets demonstrate that the proposed model is compact and yet discriminative for efficient representation of land-use scene images, and achieves competitive classification results with the state-of-art methods.

Visual attention based bag-of-words model for image classification

NASA Astrophysics Data System (ADS)

Wang, Qiwei; Wan, Shouhong; Yue, Lihua; Wang, Che

2014-04-01

Bag-of-words is a classical method for image classification. The core problem is how to count the frequency of the visual words and what visual words to select. In this paper, we propose a visual attention based bag-of-words model (VABOW model) for image classification task. The VABOW model utilizes visual attention method to generate a saliency map, and uses the saliency map as a weighted matrix to instruct the statistic process for the frequency of the visual words. On the other hand, the VABOW model combines shape, color and texture cues and uses L1 regularization logistic regression method to select the most relevant and most efficient features. We compare our approach with traditional bag-of-words based method on two datasets, and the result shows that our VABOW model outperforms the state-of-the-art method for image classification.

Cognitive Task Analysis of the HALIFAX-Class Operations Room Officer

DTIC Science & Technology

1999-03-10

Image Cover Sheet CLASSIFICATION SYSTEM NUMBER 510918 UNCLASSIFIED llllllllllllllllllllllllllllllllllllllll TITLE COGNITIVE TASK ANALYSIS OF THE...DATES COVERED 00-00-1999 to 00-00-1999 4. TITLE AND SUBTITLE Cognitive Task Analysis of the HALIFAX-Class Operations Room Officer 5a. CONTRACT...Ontario . ~ -- . ’ c ... - Incorporated Cognitive Task Analysis of the HALIFAX-Class Operations Room Officer: PWGSC Contract No. W7711-7-7404/001/SV

Boosting CNN performance for lung texture classification using connected filtering

NASA Astrophysics Data System (ADS)

Tarando, Sebastián. Roberto; Fetita, Catalin; Kim, Young-Wouk; Cho, Hyoun; Brillet, Pierre-Yves

2018-02-01

Infiltrative lung diseases describe a large group of irreversible lung disorders requiring regular follow-up with CT imaging. Quantifying the evolution of the patient status imposes the development of automated classification tools for lung texture. This paper presents an original image pre-processing framework based on locally connected filtering applied in multiresolution, which helps improving the learning process and boost the performance of CNN for lung texture classification. By removing the dense vascular network from images used by the CNN for lung classification, locally connected filters provide a better discrimination between different lung patterns and help regularizing the classification output. The approach was tested in a preliminary evaluation on a 10 patient database of various lung pathologies, showing an increase of 10% in true positive rate (on average for all the cases) with respect to the state of the art cascade of CNNs for this task.

A Coupled k-Nearest Neighbor Algorithm for Multi-Label Classification

DTIC Science & Technology

2015-05-22

classification, an image may contain several concepts simultaneously, such as beach, sunset and kangaroo . Such tasks are usually denoted as multi-label...informatics, a gene can belong to both metabolism and transcription classes; and in music categorization, a song may labeled as Mozart and sad. In the

Integrated Remote Sensing Modalities for Classification at a Legacy Test Site

NASA Astrophysics Data System (ADS)

Lee, D. J.; Anderson, D.; Craven, J.

2016-12-01

Detecting, locating, and characterizing suspected underground nuclear test sites is of interest to the worldwide nonproliferation monitoring community. Remote sensing provides both cultural and surface geological information over a large search area in a non-intrusive manner. We have characterized a legacy nuclear test site at the Nevada National Security Site (NNSS) using an aerial system based on RGB imagery, light detection and ranging, and hyperspectral imaging. We integrate these different remote sensing modalities to perform pattern recognition and classification tasks on the test site. These tasks include detecting cultural artifacts and exotic materials. We evaluate if the integration of different remote sensing modalities improves classification performance.

The Accuracy and Reliability of Crowdsource Annotations of Digital Retinal Images

PubMed Central

Mitry, Danny; Zutis, Kris; Dhillon, Baljean; Peto, Tunde; Hayat, Shabina; Khaw, Kay-Tee; Morgan, James E.; Moncur, Wendy; Trucco, Emanuele; Foster, Paul J.

2016-01-01

Purpose Crowdsourcing is based on outsourcing computationally intensive tasks to numerous individuals in the online community who have no formal training. Our aim was to develop a novel online tool designed to facilitate large-scale annotation of digital retinal images, and to assess the accuracy of crowdsource grading using this tool, comparing it to expert classification. Methods We used 100 retinal fundus photograph images with predetermined disease criteria selected by two experts from a large cohort study. The Amazon Mechanical Turk Web platform was used to drive traffic to our site so anonymous workers could perform a classification and annotation task of the fundus photographs in our dataset after a short training exercise. Three groups were assessed: masters only, nonmasters only and nonmasters with compulsory training. We calculated the sensitivity, specificity, and area under the curve (AUC) of receiver operating characteristic (ROC) plots for all classifications compared to expert grading, and used the Dice coefficient and consensus threshold to assess annotation accuracy. Results In total, we received 5389 annotations for 84 images (excluding 16 training images) in 2 weeks. A specificity and sensitivity of 71% (95% confidence interval [CI], 69%–74%) and 87% (95% CI, 86%–88%) was achieved for all classifications. The AUC in this study for all classifications combined was 0.93 (95% CI, 0.91–0.96). For image annotation, a maximal Dice coefficient (∼0.6) was achieved with a consensus threshold of 0.25. Conclusions This study supports the hypothesis that annotation of abnormalities in retinal images by ophthalmologically naive individuals is comparable to expert annotation. The highest AUC and agreement with expert annotation was achieved in the nonmasters with compulsory training group. Translational Relevance The use of crowdsourcing as a technique for retinal image analysis may be comparable to expert graders and has the potential to deliver timely, accurate, and cost-effective image analysis. PMID:27668130

The Accuracy and Reliability of Crowdsource Annotations of Digital Retinal Images.

PubMed

Mitry, Danny; Zutis, Kris; Dhillon, Baljean; Peto, Tunde; Hayat, Shabina; Khaw, Kay-Tee; Morgan, James E; Moncur, Wendy; Trucco, Emanuele; Foster, Paul J

2016-09-01

Crowdsourcing is based on outsourcing computationally intensive tasks to numerous individuals in the online community who have no formal training. Our aim was to develop a novel online tool designed to facilitate large-scale annotation of digital retinal images, and to assess the accuracy of crowdsource grading using this tool, comparing it to expert classification. We used 100 retinal fundus photograph images with predetermined disease criteria selected by two experts from a large cohort study. The Amazon Mechanical Turk Web platform was used to drive traffic to our site so anonymous workers could perform a classification and annotation task of the fundus photographs in our dataset after a short training exercise. Three groups were assessed: masters only, nonmasters only and nonmasters with compulsory training. We calculated the sensitivity, specificity, and area under the curve (AUC) of receiver operating characteristic (ROC) plots for all classifications compared to expert grading, and used the Dice coefficient and consensus threshold to assess annotation accuracy. In total, we received 5389 annotations for 84 images (excluding 16 training images) in 2 weeks. A specificity and sensitivity of 71% (95% confidence interval [CI], 69%-74%) and 87% (95% CI, 86%-88%) was achieved for all classifications. The AUC in this study for all classifications combined was 0.93 (95% CI, 0.91-0.96). For image annotation, a maximal Dice coefficient (∼0.6) was achieved with a consensus threshold of 0.25. This study supports the hypothesis that annotation of abnormalities in retinal images by ophthalmologically naive individuals is comparable to expert annotation. The highest AUC and agreement with expert annotation was achieved in the nonmasters with compulsory training group. The use of crowdsourcing as a technique for retinal image analysis may be comparable to expert graders and has the potential to deliver timely, accurate, and cost-effective image analysis.

Robust multitask learning with three-dimensional empirical mode decomposition-based features for hyperspectral classification

NASA Astrophysics Data System (ADS)

He, Zhi; Liu, Lin

2016-11-01

Empirical mode decomposition (EMD) and its variants have recently been applied for hyperspectral image (HSI) classification due to their ability to extract useful features from the original HSI. However, it remains a challenging task to effectively exploit the spectral-spatial information by the traditional vector or image-based methods. In this paper, a three-dimensional (3D) extension of EMD (3D-EMD) is proposed to naturally treat the HSI as a cube and decompose the HSI into varying oscillations (i.e. 3D intrinsic mode functions (3D-IMFs)). To achieve fast 3D-EMD implementation, 3D Delaunay triangulation (3D-DT) is utilized to determine the distances of extrema, while separable filters are adopted to generate the envelopes. Taking the extracted 3D-IMFs as features of different tasks, robust multitask learning (RMTL) is further proposed for HSI classification. In RMTL, pairs of low-rank and sparse structures are formulated by trace-norm and l1,2 -norm to capture task relatedness and specificity, respectively. Moreover, the optimization problems of RMTL can be efficiently solved by the inexact augmented Lagrangian method (IALM). Compared with several state-of-the-art feature extraction and classification methods, the experimental results conducted on three benchmark data sets demonstrate the superiority of the proposed methods.

Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images.

PubMed

Ran, Lingyan; Zhang, Yanning; Zhang, Qilin; Yang, Tao

2017-06-12

Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the "navigation via classification" task, we introduce the spherical camera for scene capturing, which enables 360° fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications.

Quantitative evaluation of variations in rule-based classifications of land cover in urban neighbourhoods using WorldView-2 imagery.

PubMed

Belgiu, Mariana; Dr Guţ, Lucian; Strobl, Josef

2014-01-01

The increasing availability of high resolution imagery has triggered the need for automated image analysis techniques, with reduced human intervention and reproducible analysis procedures. The knowledge gained in the past might be of use to achieving this goal, if systematically organized into libraries which would guide the image analysis procedure. In this study we aimed at evaluating the variability of digital classifications carried out by three experts who were all assigned the same interpretation task. Besides the three classifications performed by independent operators, we developed an additional rule-based classification that relied on the image classifications best practices found in the literature, and used it as a surrogate for libraries of object characteristics. The results showed statistically significant differences among all operators who classified the same reference imagery. The classifications carried out by the experts achieved satisfactory results when transferred to another area for extracting the same classes of interest, without modification of the developed rules.

Quantitative evaluation of variations in rule-based classifications of land cover in urban neighbourhoods using WorldView-2 imagery

PubMed Central

Belgiu, Mariana; Drǎguţ, Lucian; Strobl, Josef

2014-01-01

The increasing availability of high resolution imagery has triggered the need for automated image analysis techniques, with reduced human intervention and reproducible analysis procedures. The knowledge gained in the past might be of use to achieving this goal, if systematically organized into libraries which would guide the image analysis procedure. In this study we aimed at evaluating the variability of digital classifications carried out by three experts who were all assigned the same interpretation task. Besides the three classifications performed by independent operators, we developed an additional rule-based classification that relied on the image classifications best practices found in the literature, and used it as a surrogate for libraries of object characteristics. The results showed statistically significant differences among all operators who classified the same reference imagery. The classifications carried out by the experts achieved satisfactory results when transferred to another area for extracting the same classes of interest, without modification of the developed rules. PMID:24623959

Quantitative evaluation of variations in rule-based classifications of land cover in urban neighbourhoods using WorldView-2 imagery

NASA Astrophysics Data System (ADS)

Belgiu, Mariana; ǎguţ, Lucian, , Dr; Strobl, Josef

2014-01-01

The increasing availability of high resolution imagery has triggered the need for automated image analysis techniques, with reduced human intervention and reproducible analysis procedures. The knowledge gained in the past might be of use to achieving this goal, if systematically organized into libraries which would guide the image analysis procedure. In this study we aimed at evaluating the variability of digital classifications carried out by three experts who were all assigned the same interpretation task. Besides the three classifications performed by independent operators, we developed an additional rule-based classification that relied on the image classifications best practices found in the literature, and used it as a surrogate for libraries of object characteristics. The results showed statistically significant differences among all operators who classified the same reference imagery. The classifications carried out by the experts achieved satisfactory results when transferred to another area for extracting the same classes of interest, without modification of the developed rules.

A probabilistic approach to segmentation and classification of neoplasia in uterine cervix images using color and geometric features

NASA Astrophysics Data System (ADS)

Srinivasan, Yeshwanth; Hernes, Dana; Tulpule, Bhakti; Yang, Shuyu; Guo, Jiangling; Mitra, Sunanda; Yagneswaran, Sriraja; Nutter, Brian; Jeronimo, Jose; Phillips, Benny; Long, Rodney; Ferris, Daron

2005-04-01

Automated segmentation and classification of diagnostic markers in medical imagery are challenging tasks. Numerous algorithms for segmentation and classification based on statistical approaches of varying complexity are found in the literature. However, the design of an efficient and automated algorithm for precise classification of desired diagnostic markers is extremely image-specific. The National Library of Medicine (NLM), in collaboration with the National Cancer Institute (NCI), is creating an archive of 60,000 digitized color images of the uterine cervix. NLM is developing tools for the analysis and dissemination of these images over the Web for the study of visual features correlated with precancerous neoplasia and cancer. To enable indexing of images of the cervix, it is essential to develop algorithms for the segmentation of regions of interest, such as acetowhitened regions, and automatic identification and classification of regions exhibiting mosaicism and punctation. Success of such algorithms depends, primarily, on the selection of relevant features representing the region of interest. We present color and geometric features based statistical classification and segmentation algorithms yielding excellent identification of the regions of interest. The distinct classification of the mosaic regions from the non-mosaic ones has been obtained by clustering multiple geometric and color features of the segmented sections using various morphological and statistical approaches. Such automated classification methodologies will facilitate content-based image retrieval from the digital archive of uterine cervix and have the potential of developing an image based screening tool for cervical cancer.

Spatial Uncertainty Modeling of Fuzzy Information in Images for Pattern Classification

PubMed Central

Pham, Tuan D.

2014-01-01

The modeling of the spatial distribution of image properties is important for many pattern recognition problems in science and engineering. Mathematical methods are needed to quantify the variability of this spatial distribution based on which a decision of classification can be made in an optimal sense. However, image properties are often subject to uncertainty due to both incomplete and imprecise information. This paper presents an integrated approach for estimating the spatial uncertainty of vagueness in images using the theory of geostatistics and the calculus of probability measures of fuzzy events. Such a model for the quantification of spatial uncertainty is utilized as a new image feature extraction method, based on which classifiers can be trained to perform the task of pattern recognition. Applications of the proposed algorithm to the classification of various types of image data suggest the usefulness of the proposed uncertainty modeling technique for texture feature extraction. PMID:25157744

Comparison of classification algorithms for various methods of preprocessing radar images of the MSTAR base

NASA Astrophysics Data System (ADS)

Borodinov, A. A.; Myasnikov, V. V.

2018-04-01

The present work is devoted to comparing the accuracy of the known qualification algorithms in the task of recognizing local objects on radar images for various image preprocessing methods. Preprocessing involves speckle noise filtering and normalization of the object orientation in the image by the method of image moments and by a method based on the Hough transform. In comparison, the following classification algorithms are used: Decision tree; Support vector machine, AdaBoost, Random forest. The principal component analysis is used to reduce the dimension. The research is carried out on the objects from the base of radar images MSTAR. The paper presents the results of the conducted studies.

Efficient estimation of ideal-observer performance in classification tasks involving high-dimensional complex backgrounds

PubMed Central

Park, Subok; Clarkson, Eric

2010-01-01

The Bayesian ideal observer is optimal among all observers and sets an absolute upper bound for the performance of any observer in classification tasks [Van Trees, Detection, Estimation, and Modulation Theory, Part I (Academic, 1968).]. Therefore, the ideal observer should be used for objective image quality assessment whenever possible. However, computation of ideal-observer performance is difficult in practice because this observer requires the full description of unknown, statistical properties of high-dimensional, complex data arising in real life problems. Previously, Markov-chain Monte Carlo (MCMC) methods were developed by Kupinski et al. [J. Opt. Soc. Am. A 20, 430(2003) ] and by Park et al. [J. Opt. Soc. Am. A 24, B136 (2007) and IEEE Trans. Med. Imaging 28, 657 (2009) ] to estimate the performance of the ideal observer and the channelized ideal observer (CIO), respectively, in classification tasks involving non-Gaussian random backgrounds. However, both algorithms had the disadvantage of long computation times. We propose a fast MCMC for real-time estimation of the likelihood ratio for the CIO. Our simulation results show that our method has the potential to speed up ideal-observer performance in tasks involving complex data when efficient channels are used for the CIO. PMID:19884916

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.

Developing collaborative classifiers using an expert-based model

USGS Publications Warehouse

Mountrakis, G.; Watts, R.; Luo, L.; Wang, Jingyuan

2009-01-01

This paper presents a hierarchical, multi-stage adaptive strategy for image classification. We iteratively apply various classification methods (e.g., decision trees, neural networks), identify regions of parametric and geographic space where accuracy is low, and in these regions, test and apply alternate methods repeating the process until the entire image is classified. Currently, classifiers are evaluated through human input using an expert-based system; therefore, this paper acts as the proof of concept for collaborative classifiers. Because we decompose the problem into smaller, more manageable sub-tasks, our classification exhibits increased flexibility compared to existing methods since classification methods are tailored to the idiosyncrasies of specific regions. A major benefit of our approach is its scalability and collaborative support since selected low-accuracy classifiers can be easily replaced with others without affecting classification accuracy in high accuracy areas. At each stage, we develop spatially explicit accuracy metrics that provide straightforward assessment of results by non-experts and point to areas that need algorithmic improvement or ancillary data. Our approach is demonstrated in the task of detecting impervious surface areas, an important indicator for human-induced alterations to the environment, using a 2001 Landsat scene from Las Vegas, Nevada. ?? 2009 American Society for Photogrammetry and Remote Sensing.

Integrating Human and Machine Intelligence in Galaxy Morphology Classification Tasks

NASA Astrophysics Data System (ADS)

Beck, Melanie Renee

The large flood of data flowing from observatories presents significant challenges to astronomy and cosmology--challenges that will only be magnified by projects currently under development. Growth in both volume and velocity of astrophysics data is accelerating: whereas the Sloan Digital Sky Survey (SDSS) has produced 60 terabytes of data in the last decade, the upcoming Large Synoptic Survey Telescope (LSST) plans to register 30 terabytes per night starting in the year 2020. Additionally, the Euclid Mission will acquire imaging for 5 x 107 resolvable galaxies. The field of galaxy evolution faces a particularly challenging future as complete understanding often cannot be reached without analysis of detailed morphological galaxy features. Historically, morphological analysis has relied on visual classification by astronomers, accessing the human brains capacity for advanced pattern recognition. However, this accurate but inefficient method falters when confronted with many thousands (or millions) of images. In the SDSS era, efforts to automate morphological classifications of galaxies (e.g., Conselice et al., 2000; Lotz et al., 2004) are reasonably successful and can distinguish between elliptical and disk-dominated galaxies with accuracies of 80%. While this is statistically very useful, a key problem with these methods is that they often cannot say which 80% of their samples are accurate. Furthermore, when confronted with the more complex task of identifying key substructure within galaxies, automated classification algorithms begin to fail. The Galaxy Zoo project uses a highly innovative approach to solving the scalability problem of visual classification. Displaying images of SDSS galaxies to volunteers via a simple and engaging web interface, www.galaxyzoo.org asks people to classify images by eye. Within the first year hundreds of thousands of members of the general public had classified each of the 1 million SDSS galaxies an average of 40 times. Galaxy Zoo thus solved both the visual classification problem of time efficiency and improved accuracy by producing a distribution of independent classifications for each galaxy. While crowd-sourced galaxy classifications have proven their worth, challenges remain before establishing this method as a critical and standard component of the data processing pipelines for the next generation of surveys. In particular, though innovative, crowd-sourcing techniques do not have the capacity to handle the data volume and rates expected in the next generation of surveys. These algorithms will be delegated to handle the majority of the classification tasks, freeing citizen scientists to contribute their efforts on subtler and more complex assignments. This thesis presents a solution through an integration of visual and automated classifications, preserving the best features of both human and machine. We demonstrate the effectiveness of such a system through a re-analysis of visual galaxy morphology classifications collected during the Galaxy Zoo 2 (GZ2) project. We reprocess the top-level question of the GZ2 decision tree with a Bayesian classification aggregation algorithm dubbed SWAP, originally developed for the Space Warps gravitational lens project. Through a simple binary classification scheme we increase the classification rate nearly 5-fold classifying 226,124 galaxies in 92 days of GZ2 project time while reproducing labels derived from GZ2 classification data with 95.7% accuracy. We next combine this with a Random Forest machine learning algorithm that learns on a suite of non-parametric morphology indicators widely used for automated morphologies. We develop a decision engine that delegates tasks between human and machine and demonstrate that the combined system provides a factor of 11.4 increase in the classification rate, classifying 210,803 galaxies in just 32 days of GZ2 project time with 93.1% accuracy. As the Random Forest algorithm requires a minimal amount of computational cost, this result has important implications for galaxy morphology identification tasks in the era of Euclid and other large-scale surveys.

Effective Sequential Classifier Training for SVM-Based Multitemporal Remote Sensing Image Classification

NASA Astrophysics Data System (ADS)

Guo, Yiqing; Jia, Xiuping; Paull, David

2018-06-01

The explosive availability of remote sensing images has challenged supervised classification algorithms such as Support Vector Machines (SVM), as training samples tend to be highly limited due to the expensive and laborious task of ground truthing. The temporal correlation and spectral similarity between multitemporal images have opened up an opportunity to alleviate this problem. In this study, a SVM-based Sequential Classifier Training (SCT-SVM) approach is proposed for multitemporal remote sensing image classification. The approach leverages the classifiers of previous images to reduce the required number of training samples for the classifier training of an incoming image. For each incoming image, a rough classifier is firstly predicted based on the temporal trend of a set of previous classifiers. The predicted classifier is then fine-tuned into a more accurate position with current training samples. This approach can be applied progressively to sequential image data, with only a small number of training samples being required from each image. Experiments were conducted with Sentinel-2A multitemporal data over an agricultural area in Australia. Results showed that the proposed SCT-SVM achieved better classification accuracies compared with two state-of-the-art model transfer algorithms. When training data are insufficient, the overall classification accuracy of the incoming image was improved from 76.18% to 94.02% with the proposed SCT-SVM, compared with those obtained without the assistance from previous images. These results demonstrate that the leverage of a priori information from previous images can provide advantageous assistance for later images in multitemporal image classification.

Can Distributed Volunteers Accomplish Massive Data Analysis Tasks?

NASA Technical Reports Server (NTRS)

Kanefsky, B.; Barlow, N. G.; Gulick, V. C.

2001-01-01

We argue that many image analysis tasks can be performed by distributed amateurs. Our pilot study, with crater surveying and classification, has produced encouraging results in terms of both quantity (100,000 crater entries in 2 months) and quality. Additional information is contained in the original extended abstract.

Comparing fusion techniques for the ImageCLEF 2013 medical case retrieval task.

PubMed

G Seco de Herrera, Alba; Schaer, Roger; Markonis, Dimitrios; Müller, Henning

2015-01-01

Retrieval systems can supply similar cases with a proven diagnosis to a new example case under observation to help clinicians during their work. The ImageCLEFmed evaluation campaign proposes a framework where research groups can compare case-based retrieval approaches. This paper focuses on the case-based task and adds results of the compound figure separation and modality classification tasks. Several fusion approaches are compared to identify the approaches best adapted to the heterogeneous data of the task. Fusion of visual and textual features is analyzed, demonstrating that the selection of the fusion strategy can improve the best performance on the case-based retrieval task. Copyright © 2014 Elsevier Ltd. All rights reserved.

A novel underwater dam crack detection and classification approach based on sonar images

PubMed Central

Shi, Pengfei; Fan, Xinnan; Ni, Jianjun; Khan, Zubair; Li, Min

2017-01-01

Underwater dam crack detection and classification based on sonar images is a challenging task because underwater environments are complex and because cracks are quite random and diverse in nature. Furthermore, obtainable sonar images are of low resolution. To address these problems, a novel underwater dam crack detection and classification approach based on sonar imagery is proposed. First, the sonar images are divided into image blocks. Second, a clustering analysis of a 3-D feature space is used to obtain the crack fragments. Third, the crack fragments are connected using an improved tensor voting method. Fourth, a minimum spanning tree is used to obtain the crack curve. Finally, an improved evidence theory combined with fuzzy rule reasoning is proposed to classify the cracks. Experimental results show that the proposed approach is able to detect underwater dam cracks and classify them accurately and effectively under complex underwater environments. PMID:28640925

A novel underwater dam crack detection and classification approach based on sonar images.

PubMed

Shi, Pengfei; Fan, Xinnan; Ni, Jianjun; Khan, Zubair; Li, Min

2017-01-01

Underwater dam crack detection and classification based on sonar images is a challenging task because underwater environments are complex and because cracks are quite random and diverse in nature. Furthermore, obtainable sonar images are of low resolution. To address these problems, a novel underwater dam crack detection and classification approach based on sonar imagery is proposed. First, the sonar images are divided into image blocks. Second, a clustering analysis of a 3-D feature space is used to obtain the crack fragments. Third, the crack fragments are connected using an improved tensor voting method. Fourth, a minimum spanning tree is used to obtain the crack curve. Finally, an improved evidence theory combined with fuzzy rule reasoning is proposed to classify the cracks. Experimental results show that the proposed approach is able to detect underwater dam cracks and classify them accurately and effectively under complex underwater environments.

Cascade of convolutional neural networks for lung texture classification: overcoming ontological overlapping

NASA Astrophysics Data System (ADS)

Tarando, Sebastian Roberto; Fetita, Catalin; Brillet, Pierre-Yves

2017-03-01

The infiltrative lung diseases are a class of irreversible, non-neoplastic lung pathologies requiring regular follow-up with CT imaging. Quantifying the evolution of the patient status imposes the development of automated classification tools for lung texture. Traditionally, such classification relies on a two-dimensional analysis of axial CT images. This paper proposes a cascade of the existing CNN based CAD system, specifically tuned-up. The advantage of using a deep learning approach is a better regularization of the classification output. In a preliminary evaluation, the combined approach was tested on a 13 patient database of various lung pathologies, showing an increase of 10% in True Positive Rate (TPR) with respect to the best suited state of the art CNN for this task.

Classification images reveal decision variables and strategies in forced choice tasks

PubMed Central

Pritchett, Lisa M.; Murray, Richard F.

2015-01-01

Despite decades of research, there is still uncertainty about how people make simple decisions about perceptual stimuli. Most theories assume that perceptual decisions are based on decision variables, which are internal variables that encode task-relevant information. However, decision variables are usually considered to be theoretical constructs that cannot be measured directly, and this often makes it difficult to test theories of perceptual decision making. Here we show how to measure decision variables on individual trials, and we use these measurements to test theories of perceptual decision making more directly than has previously been possible. We measure classification images, which are estimates of templates that observers use to extract information from stimuli. We then calculate the dot product of these classification images with the stimuli to estimate observers' decision variables. Finally, we reconstruct each observer's “decision space,” a map that shows the probability of the observer’s responses for all values of the decision variables. We use this method to examine decision strategies in two-alternative forced choice (2AFC) tasks, for which there are several competing models. In one experiment, the resulting decision spaces support the difference model, a classic theory of 2AFC decisions. In a second experiment, we find unexpected decision spaces that are not predicted by standard models of 2AFC decisions, and that suggest intrinsic uncertainty or soft thresholding. These experiments give new evidence regarding observers’ strategies in 2AFC tasks, and they show how measuring decision variables can answer long-standing questions about perceptual decision making. PMID:26015584

Assessment of CT image quality using a Bayesian approach

NASA Astrophysics Data System (ADS)

Reginatto, M.; Anton, M.; Elster, C.

2017-08-01

One of the most promising approaches for evaluating CT image quality is task-specific quality assessment. This involves a simplified version of a clinical task, e.g. deciding whether an image belongs to the class of images that contain the signature of a lesion or not. Task-specific quality assessment can be done by model observers, which are mathematical procedures that carry out the classification task. The most widely used figure of merit for CT image quality is the area under the ROC curve (AUC), a quantity which characterizes the performance of a given model observer. In order to estimate AUC from a finite sample of images, different approaches from classical statistics have been suggested. The goal of this paper is to introduce task-specific quality assessment of CT images to metrology and to propose a novel Bayesian estimation of AUC for the channelized Hotelling observer (CHO) applied to the task of detecting a lesion at a known image location. It is assumed that signal-present and signal-absent images follow multivariate normal distributions with the same covariance matrix. The Bayesian approach results in a posterior distribution for the AUC of the CHO which provides in addition a complete characterization of the uncertainty of this figure of merit. The approach is illustrated by its application to both simulated and experimental data.

An integration of minimum local feature representation methods to recognize large variation of foods

NASA Astrophysics Data System (ADS)

Razali, Mohd Norhisham bin; Manshor, Noridayu; Halin, Alfian Abdul; Mustapha, Norwati; Yaakob, Razali

2017-10-01

Local invariant features have shown to be successful in describing object appearances for image classification tasks. Such features are robust towards occlusion and clutter and are also invariant against scale and orientation changes. This makes them suitable for classification tasks with little inter-class similarity and large intra-class difference. In this paper, we propose an integrated representation of the Speeded-Up Robust Feature (SURF) and Scale Invariant Feature Transform (SIFT) descriptors, using late fusion strategy. The proposed representation is used for food recognition from a dataset of food images with complex appearance variations. The Bag of Features (BOF) approach is employed to enhance the discriminative ability of the local features. Firstly, the individual local features are extracted to construct two kinds of visual vocabularies, representing SURF and SIFT. The visual vocabularies are then concatenated and fed into a Linear Support Vector Machine (SVM) to classify the respective food categories. Experimental results demonstrate impressive overall recognition at 82.38% classification accuracy based on the challenging UEC-Food100 dataset.

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.

Multi-Temporal Classification and Change Detection Using Uav Images

NASA Astrophysics Data System (ADS)

Makuti, S.; Nex, F.; Yang, M. Y.

2018-05-01

In this paper different methodologies for the classification and change detection of UAV image blocks are explored. UAV is not only the cheapest platform for image acquisition but it is also the easiest platform to operate in repeated data collections over a changing area like a building construction site. Two change detection techniques have been evaluated in this study: the pre-classification and the post-classification algorithms. These methods are based on three main steps: feature extraction, classification and change detection. A set of state of the art features have been used in the tests: colour features (HSV), textural features (GLCM) and 3D geometric features. For classification purposes Conditional Random Field (CRF) has been used: the unary potential was determined using the Random Forest algorithm while the pairwise potential was defined by the fully connected CRF. In the performed tests, different feature configurations and settings have been considered to assess the performance of these methods in such challenging task. Experimental results showed that the post-classification approach outperforms the pre-classification change detection method. This was analysed using the overall accuracy, where by post classification have an accuracy of up to 62.6 % and the pre classification change detection have an accuracy of 46.5 %. These results represent a first useful indication for future works and developments.

Tiny videos: a large data set for nonparametric video retrieval and frame classification.

PubMed

Karpenko, Alexandre; Aarabi, Parham

2011-03-01

In this paper, we present a large database of over 50,000 user-labeled videos collected from YouTube. We develop a compact representation called "tiny videos" that achieves high video compression rates while retaining the overall visual appearance of the video as it varies over time. We show that frame sampling using affinity propagation-an exemplar-based clustering algorithm-achieves the best trade-off between compression and video recall. We use this large collection of user-labeled videos in conjunction with simple data mining techniques to perform related video retrieval, as well as classification of images and video frames. The classification results achieved by tiny videos are compared with the tiny images framework [24] for a variety of recognition tasks. The tiny images data set consists of 80 million images collected from the Internet. These are the largest labeled research data sets of videos and images available to date. We show that tiny videos are better suited for classifying scenery and sports activities, while tiny images perform better at recognizing objects. Furthermore, we demonstrate that combining the tiny images and tiny videos data sets improves classification precision in a wider range of categories.

Observer efficiency in discrimination tasks simulating malignant and benign breast lesions imaged with ultrasound

PubMed Central

Abbey, Craig K.; Zemp, Roger J.; Liu, Jie; Lindfors, Karen K.; Insana, Michael F.

2009-01-01

We investigate and extend the ideal observer methodology developed by Smith and Wagner to detection and discrimination tasks related to breast sonography. We provide a numerical approach for evaluating the ideal observer acting on radio-frequency (RF) frame data, which involves inversion of large nonstationary covariance matrices, and we describe a power-series approach to computing this inverse. Considering a truncated power series suggests that the RF data be Wiener-filtered before forming the final envelope image. We have compared human performance for Wiener-filtered and conventional B-mode envelope images using psychophysical studies for 5 tasks related to breast cancer classification. We find significant improvements in visual detection and discrimination efficiency in four of these five tasks. We also use the Smith-Wagner approach to distinguish between human and processing inefficiencies, and find that generally the principle limitation comes from the information lost in computing the final envelope image. PMID:16468454

Lung texture classification using bag of visual words

NASA Astrophysics Data System (ADS)

Asherov, Marina; Diamant, Idit; Greenspan, Hayit

2014-03-01

Interstitial lung diseases (ILD) refer to a group of more than 150 parenchymal lung disorders. High-Resolution Computed Tomography (HRCT) is the most essential imaging modality of ILD diagnosis. Nonetheless, classification of various lung tissue patterns caused by ILD is still regarded as a challenging task. The current study focuses on the classification of five most common categories of lung tissues of ILD in HRCT images: normal, emphysema, ground glass, fibrosis and micronodules. The objective of the research is to classify an expert-given annotated region of interest (AROI) using a bag of visual words (BoVW) framework. The images are divided into small patches and a collection of representative patches are defined as visual words. This procedure, termed dictionary construction, is performed for each individual lung texture category. The assumption is that different lung textures are represented by a different visual word distribution. The classification is performed using an SVM classifier with histogram intersection kernel. In the experiments, we use a dataset of 1018 AROIs from 95 patients. Classification using a leave-one-patient-out cross validation (LOPO CV) is used. Current classification accuracy obtained is close to 80%.

Integration of heterogeneous features for remote sensing scene classification

NASA Astrophysics Data System (ADS)

Wang, Xin; Xiong, Xingnan; Ning, Chen; Shi, Aiye; Lv, Guofang

2018-01-01

Scene classification is one of the most important issues in remote sensing (RS) image processing. We find that features from different channels (shape, spectral, texture, etc.), levels (low-level and middle-level), or perspectives (local and global) could provide various properties for RS images, and then propose a heterogeneous feature framework to extract and integrate heterogeneous features with different types for RS scene classification. The proposed method is composed of three modules (1) heterogeneous features extraction, where three heterogeneous feature types, called DS-SURF-LLC, mean-Std-LLC, and MS-CLBP, are calculated, (2) heterogeneous features fusion, where the multiple kernel learning (MKL) is utilized to integrate the heterogeneous features, and (3) an MKL support vector machine classifier for RS scene classification. The proposed method is extensively evaluated on three challenging benchmark datasets (a 6-class dataset, a 12-class dataset, and a 21-class dataset), and the experimental results show that the proposed method leads to good classification performance. It produces good informative features to describe the RS image scenes. Moreover, the integration of heterogeneous features outperforms some state-of-the-art features on RS scene classification tasks.

Cognitive Task Analysis of the HALIFAX-Class Operations Room Officer: Data Sheets. Annexes

DTIC Science & Technology

1999-03-10

Image Cover Sheet CLASSIFICATION SYSTEM NUMBER 510920 UNCLASSIFIED 1111111111111111111111111111111111111111 TITLE ANNEXES TO: COGNITIVE TASK ANALYSIS OF...1999 2. REPORT TYPE 3. DATES COVERED 00-00-1999 to 00-00-1999 4. TITLE AND SUBTITLE Annexes to: Cognitive Task Analysis of the HALIFAX-Class...by ANSI Std Z39-18 Guelph, Ontario .H U. M A N S X S T E M S Incorporated Annexes to: Cognitive Task Analysis of the HALIFAX-Class Operations

The Ilac-Project Supporting Ancient Coin Classification by Means of Image Analysis

NASA Astrophysics Data System (ADS)

Kavelar, A.; Zambanini, S.; Kampel, M.; Vondrovec, K.; Siegl, K.

2013-07-01

This paper presents the ILAC project, which aims at the development of an automated image-based classification system for ancient Roman Republican coins. The benefits of such a system are manifold: operating at the suture between computer vision and numismatics, ILAC can reduce the day-to-day workload of numismatists by assisting them in classification tasks and providing a preselection of suitable coin classes. This is especially helpful for large coin hoard findings comprising several thousands of coins. Furthermore, this system could be implemented in an online platform for hobby numismatists, allowing them to access background information about their coin collection by simply uploading a photo of obverse and reverse for the coin of interest. ILAC explores different computer vision techniques and their combinations for the use of image-based coin recognition. Some of these methods, such as image matching, use the entire coin image in the classification process, while symbol or legend recognition exploit certain characteristics of the coin imagery. An overview of the methods explored so far and the respective experiments is given as well as an outlook on the next steps of the project.

Automatic Screening and Grading of Age-Related Macular Degeneration from Texture Analysis of Fundus Images

PubMed Central

Phan, Thanh Vân; Seoud, Lama; Chakor, Hadi; Cheriet, Farida

2016-01-01

Age-related macular degeneration (AMD) is a disease which causes visual deficiency and irreversible blindness to the elderly. In this paper, an automatic classification method for AMD is proposed to perform robust and reproducible assessments in a telemedicine context. First, a study was carried out to highlight the most relevant features for AMD characterization based on texture, color, and visual context in fundus images. A support vector machine and a random forest were used to classify images according to the different AMD stages following the AREDS protocol and to evaluate the features' relevance. Experiments were conducted on a database of 279 fundus images coming from a telemedicine platform. The results demonstrate that local binary patterns in multiresolution are the most relevant for AMD classification, regardless of the classifier used. Depending on the classification task, our method achieves promising performances with areas under the ROC curve between 0.739 and 0.874 for screening and between 0.469 and 0.685 for grading. Moreover, the proposed automatic AMD classification system is robust with respect to image quality. PMID:27190636

Patch-based Convolutional Neural Network for Whole Slide Tissue Image Classification

PubMed Central

Hou, Le; Samaras, Dimitris; Kurc, Tahsin M.; Gao, Yi; Davis, James E.; Saltz, Joel H.

2016-01-01

Convolutional Neural Networks (CNN) are state-of-the-art models for many image classification tasks. However, to recognize cancer subtypes automatically, training a CNN on gigapixel resolution Whole Slide Tissue Images (WSI) is currently computationally impossible. The differentiation of cancer subtypes is based on cellular-level visual features observed on image patch scale. Therefore, we argue that in this situation, training a patch-level classifier on image patches will perform better than or similar to an image-level classifier. The challenge becomes how to intelligently combine patch-level classification results and model the fact that not all patches will be discriminative. We propose to train a decision fusion model to aggregate patch-level predictions given by patch-level CNNs, which to the best of our knowledge has not been shown before. Furthermore, we formulate a novel Expectation-Maximization (EM) based method that automatically locates discriminative patches robustly by utilizing the spatial relationships of patches. We apply our method to the classification of glioma and non-small-cell lung carcinoma cases into subtypes. The classification accuracy of our method is similar to the inter-observer agreement between pathologists. Although it is impossible to train CNNs on WSIs, we experimentally demonstrate using a comparable non-cancer dataset of smaller images that a patch-based CNN can outperform an image-based CNN. PMID:27795661

Contextual convolutional neural networks for lung nodule classification using Gaussian-weighted average image patches

NASA Astrophysics Data System (ADS)

Lee, Haeil; Lee, Hansang; Park, Minseok; Kim, Junmo

2017-03-01

Lung cancer is the most common cause of cancer-related death. To diagnose lung cancers in early stages, numerous studies and approaches have been developed for cancer screening with computed tomography (CT) imaging. In recent years, convolutional neural networks (CNN) have become one of the most common and reliable techniques in computer aided detection (CADe) and diagnosis (CADx) by achieving state-of-the-art-level performances for various tasks. In this study, we propose a CNN classification system for false positive reduction of initially detected lung nodule candidates. First, image patches of lung nodule candidates are extracted from CT scans to train a CNN classifier. To reflect the volumetric contextual information of lung nodules to 2D image patch, we propose a weighted average image patch (WAIP) generation by averaging multiple slice images of lung nodule candidates. Moreover, to emphasize central slices of lung nodules, slice images are locally weighted according to Gaussian distribution and averaged to generate the 2D WAIP. With these extracted patches, 2D CNN is trained to achieve the classification of WAIPs of lung nodule candidates into positive and negative labels. We used LUNA 2016 public challenge database to validate the performance of our approach for false positive reduction in lung CT nodule classification. Experiments show our approach improves the classification accuracy of lung nodules compared to the baseline 2D CNN with patches from single slice image.

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.

Binary Image Classification: A Genetic Programming Approach to the Problem of Limited Training Instances.

PubMed

Al-Sahaf, Harith; Zhang, Mengjie; Johnston, Mark

2016-01-01

In the computer vision and pattern recognition fields, image classification represents an important yet difficult task. It is a challenge to build effective computer models to replicate the remarkable ability of the human visual system, which relies on only one or a few instances to learn a completely new class or an object of a class. Recently we proposed two genetic programming (GP) methods, one-shot GP and compound-GP, that aim to evolve a program for the task of binary classification in images. The two methods are designed to use only one or a few instances per class to evolve the model. In this study, we investigate these two methods in terms of performance, robustness, and complexity of the evolved programs. We use ten data sets that vary in difficulty to evaluate these two methods. We also compare them with two other GP and six non-GP methods. The results show that one-shot GP and compound-GP outperform or achieve results comparable to competitor methods. Moreover, the features extracted by these two methods improve the performance of other classifiers with handcrafted features and those extracted by a recently developed GP-based method in most cases.

MO-DE-207A-02: A Feature-Preserving Image Reconstruction Method for Improved Pancreaticlesion Classification in Diagnostic CT Imaging

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

Xu, J; Tsui, B; Noo, F

Purpose: To develop a feature-preserving model based image reconstruction (MBIR) method that improves performance in pancreatic lesion classification at equal or reduced radiation dose. Methods: A set of pancreatic lesion models was created with both benign and premalignant lesion types. These two classes of lesions are distinguished by their fine internal structures; their delineation is therefore crucial to the task of pancreatic lesion classification. To reduce image noise while preserving the features of the lesions, we developed a MBIR method with curvature-based regularization. The novel regularization encourages formation of smooth surfaces that model both the exterior shape and the internalmore » features of pancreatic lesions. Given that the curvature depends on the unknown image, image reconstruction or denoising becomes a non-convex optimization problem; to address this issue an iterative-reweighting scheme was used to calculate and update the curvature using the image from the previous iteration. Evaluation was carried out with insertion of the lesion models into the pancreas of a patient CT image. Results: Visual inspection was used to compare conventional TV regularization with our curvature-based regularization. Several penalty-strengths were considered for TV regularization, all of which resulted in erasing portions of the septation (thin partition) in a premalignant lesion. At matched noise variance (50% noise reduction in the patient stomach region), the connectivity of the septation was well preserved using the proposed curvature-based method. Conclusion: The curvature-based regularization is able to reduce image noise while simultaneously preserving the lesion features. This method could potentially improve task performance for pancreatic lesion classification at equal or reduced radiation dose. The result is of high significance for longitudinal surveillance studies of patients with pancreatic cysts, which may develop into pancreatic cancer. The Senior Author receives financial support from Siemens GmbH Healthcare.« less

Parallel-hierarchical processing and classification of laser beam profile images based on the GPU-oriented architecture

NASA Astrophysics Data System (ADS)

Yarovyi, Andrii A.; Timchenko, Leonid I.; Kozhemiako, Volodymyr P.; Kokriatskaia, Nataliya I.; Hamdi, Rami R.; Savchuk, Tamara O.; Kulyk, Oleksandr O.; Surtel, Wojciech; Amirgaliyev, Yedilkhan; Kashaganova, Gulzhan

2017-08-01

The paper deals with a problem of insufficient productivity of existing computer means for large image processing, which do not meet modern requirements posed by resource-intensive computing tasks of laser beam profiling. The research concentrated on one of the profiling problems, namely, real-time processing of spot images of the laser beam profile. Development of a theory of parallel-hierarchic transformation allowed to produce models for high-performance parallel-hierarchical processes, as well as algorithms and software for their implementation based on the GPU-oriented architecture using GPGPU technologies. The analyzed performance of suggested computerized tools for processing and classification of laser beam profile images allows to perform real-time processing of dynamic images of various sizes.

A new blood vessel extraction technique using edge enhancement and object classification.

PubMed

Badsha, Shahriar; Reza, Ahmed Wasif; Tan, Kim Geok; Dimyati, Kaharudin

2013-12-01

Diabetic retinopathy (DR) is increasing progressively pushing the demand of automatic extraction and classification of severity of diseases. Blood vessel extraction from the fundus image is a vital and challenging task. Therefore, this paper presents a new, computationally simple, and automatic method to extract the retinal blood vessel. The proposed method comprises several basic image processing techniques, namely edge enhancement by standard template, noise removal, thresholding, morphological operation, and object classification. The proposed method has been tested on a set of retinal images. The retinal images were collected from the DRIVE database and we have employed robust performance analysis to evaluate the accuracy. The results obtained from this study reveal that the proposed method offers an average accuracy of about 97 %, sensitivity of 99 %, specificity of 86 %, and predictive value of 98 %, which is superior to various well-known techniques.

Locality constrained joint dynamic sparse representation for local matching based face recognition.

PubMed

Wang, Jianzhong; Yi, Yugen; Zhou, Wei; Shi, Yanjiao; Qi, Miao; Zhang, Ming; Zhang, Baoxue; Kong, Jun

2014-01-01

Recently, Sparse Representation-based Classification (SRC) has attracted a lot of attention for its applications to various tasks, especially in biometric techniques such as face recognition. However, factors such as lighting, expression, pose and disguise variations in face images will decrease the performances of SRC and most other face recognition techniques. In order to overcome these limitations, we propose a robust face recognition method named Locality Constrained Joint Dynamic Sparse Representation-based Classification (LCJDSRC) in this paper. In our method, a face image is first partitioned into several smaller sub-images. Then, these sub-images are sparsely represented using the proposed locality constrained joint dynamic sparse representation algorithm. Finally, the representation results for all sub-images are aggregated to obtain the final recognition result. Compared with other algorithms which process each sub-image of a face image independently, the proposed algorithm regards the local matching-based face recognition as a multi-task learning problem. Thus, the latent relationships among the sub-images from the same face image are taken into account. Meanwhile, the locality information of the data is also considered in our algorithm. We evaluate our algorithm by comparing it with other state-of-the-art approaches. Extensive experiments on four benchmark face databases (ORL, Extended YaleB, AR and LFW) demonstrate the effectiveness of LCJDSRC.

Plant species classification using flower images—A comparative study of local feature representations

PubMed Central

Seeland, Marco; Rzanny, Michael; Alaqraa, Nedal; Wäldchen, Jana; Mäder, Patrick

2017-01-01

Steady improvements of image description methods induced a growing interest in image-based plant species classification, a task vital to the study of biodiversity and ecological sensitivity. Various techniques have been proposed for general object classification over the past years and several of them have already been studied for plant species classification. However, results of these studies are selective in the evaluated steps of a classification pipeline, in the utilized datasets for evaluation, and in the compared baseline methods. No study is available that evaluates the main competing methods for building an image representation on the same datasets allowing for generalized findings regarding flower-based plant species classification. The aim of this paper is to comparatively evaluate methods, method combinations, and their parameters towards classification accuracy. The investigated methods span from detection, extraction, fusion, pooling, to encoding of local features for quantifying shape and color information of flower images. We selected the flower image datasets Oxford Flower 17 and Oxford Flower 102 as well as our own Jena Flower 30 dataset for our experiments. Findings show large differences among the various studied techniques and that their wisely chosen orchestration allows for high accuracies in species classification. We further found that true local feature detectors in combination with advanced encoding methods yield higher classification results at lower computational costs compared to commonly used dense sampling and spatial pooling methods. Color was found to be an indispensable feature for high classification results, especially while preserving spatial correspondence to gray-level features. In result, our study provides a comprehensive overview of competing techniques and the implications of their main parameters for flower-based plant species classification. PMID:28234999

The highs and lows of object impossibility: effects of spatial frequency on holistic processing of impossible objects.

PubMed

Freud, Erez; Avidan, Galia; Ganel, Tzvi

2015-02-01

Holistic processing, the decoding of a stimulus as a unified whole, is a basic characteristic of object perception. Recent research using Garner's speeded classification task has shown that this processing style is utilized even for impossible objects that contain an inherent spatial ambiguity. In particular, similar Garner interference effects were found for possible and impossible objects, indicating similar holistic processing styles for the two object categories. In the present study, we further investigated the perceptual mechanisms that mediate such holistic representation of impossible objects. We relied on the notion that, whereas information embedded in the high-spatial-frequency (HSF) content supports fine-detailed processing of object features, the information conveyed by low spatial frequencies (LSF) is more crucial for the emergence of a holistic shape representation. To test the effects of image frequency on the holistic processing of impossible objects, participants performed the Garner speeded classification task on images of possible and impossible cubes filtered for their LSF and HSF information. For images containing only LSF, similar interference effects were observed for possible and impossible objects, indicating that the two object categories were processed in a holistic manner. In contrast, for the HSF images, Garner interference was obtained only for possible, but not for impossible objects. Importantly, we provided evidence to show that this effect could not be attributed to a lack of sensitivity to object possibility in the LSF images. Particularly, even for full-spectrum images, Garner interference was still observed for both possible and impossible objects. Additionally, performance in an object classification task revealed high sensitivity to object possibility, even for LSF images. Taken together, these findings suggest that the visual system can tolerate the spatial ambiguity typical to impossible objects by relying on information embedded in LSF, whereas HSF information may underlie the visual system's susceptibility to distortions in objects' spatial layouts.

Spatially aggregated multiclass pattern classification in functional MRI using optimally selected functional brain areas.

PubMed

Zheng, Weili; Ackley, Elena S; Martínez-Ramón, Manel; Posse, Stefan

2013-02-01

In previous works, boosting aggregation of classifier outputs from discrete brain areas has been demonstrated to reduce dimensionality and improve the robustness and accuracy of functional magnetic resonance imaging (fMRI) classification. However, dimensionality reduction and classification of mixed activation patterns of multiple classes remain challenging. In the present study, the goals were (a) to reduce dimensionality by combining feature reduction at the voxel level and backward elimination of optimally aggregated classifiers at the region level, (b) to compare region selection for spatially aggregated classification using boosting and partial least squares regression methods and (c) to resolve mixed activation patterns using probabilistic prediction of individual tasks. Brain activation maps from interleaved visual, motor, auditory and cognitive tasks were segmented into 144 functional regions. Feature selection reduced the number of feature voxels by more than 50%, leaving 95 regions. The two aggregation approaches further reduced the number of regions to 30, resulting in more than 75% reduction of classification time and misclassification rates of less than 3%. Boosting and partial least squares (PLS) were compared to select the most discriminative and the most task correlated regions, respectively. Successful task prediction in mixed activation patterns was feasible within the first block of task activation in real-time fMRI experiments. This methodology is suitable for sparsifying activation patterns in real-time fMRI and for neurofeedback from distributed networks of brain activation. Copyright © 2013 Elsevier Inc. All rights reserved.

Novel Algorithm for Classification of Medical Images

NASA Astrophysics Data System (ADS)

Bhushan, Bharat; Juneja, Monika

2010-11-01

Content-based image retrieval (CBIR) methods in medical image databases have been designed to support specific tasks, such as retrieval of medical images. These methods cannot be transferred to other medical applications since different imaging modalities require different types of processing. To enable content-based queries in diverse collections of medical images, the retrieval system must be familiar with the current Image class prior to the query processing. Further, almost all of them deal with the DICOM imaging format. In this paper a novel algorithm based on energy information obtained from wavelet transform for the classification of medical images according to their modalities is described. For this two types of wavelets have been used and have been shown that energy obtained in either case is quite distinct for each of the body part. This technique can be successfully applied to different image formats. The results are shown for JPEG imaging format.

A domain-knowledge-inspired mathematical framework for the description and classification of H&E stained histopathology images.

PubMed

Massar, Melody L; Bhagavatula, Ramamurthy; Ozolek, John A; Castro, Carlos A; Fickus, Matthew; Kovačević, Jelena

2011-10-19

We present the current state of our work on a mathematical framework for identification and delineation of histopathology images-local histograms and occlusion models. Local histograms are histograms computed over defined spatial neighborhoods whose purpose is to characterize an image locally. This unit of description is augmented by our occlusion models that describe a methodology for image formation. In the context of this image formation model, the power of local histograms with respect to appropriate families of images will be shown through various proved statements about expected performance. We conclude by presenting a preliminary study to demonstrate the power of the framework in the context of histopathology image classification tasks that, while differing greatly in application, both originate from what is considered an appropriate class of images for this framework.

Employing wavelet-based texture features in ammunition classification

NASA Astrophysics Data System (ADS)

Borzino, Ángelo M. C. R.; Maher, Robert C.; Apolinário, José A.; de Campos, Marcello L. R.

2017-05-01

Pattern recognition, a branch of machine learning, involves classification of information in images, sounds, and other digital representations. This paper uses pattern recognition to identify which kind of ammunition was used when a bullet was fired based on a carefully constructed set of gunshot sound recordings. To do this task, we show that texture features obtained from the wavelet transform of a component of the gunshot signal, treated as an image, and quantized in gray levels, are good ammunition discriminators. We test the technique with eight different calibers and achieve a classification rate better than 95%. We also compare the performance of the proposed method with results obtained by standard temporal and spectrographic techniques

Automated detection and classification of dice

NASA Astrophysics Data System (ADS)

Correia, Bento A. B.; Silva, Jeronimo A.; Carvalho, Fernando D.; Guilherme, Rui; Rodrigues, Fernando C.; de Silva Ferreira, Antonio M.

1995-03-01

This paper describes a typical machine vision system in an unusual application, the automated visual inspection of a Casino's playing tables. The SORTE computer vision system was developed at INETI under a contract with the Portuguese Gaming Inspection Authorities IGJ. It aims to automate the tasks of detection and classification of the dice's scores on the playing tables of the game `Banca Francesa' (which means French Banking) in Casinos. The system is based on the on-line analysis of the images captured by a monochrome CCD camera placed over the playing tables, in order to extract relevant information concerning the score indicated by the dice. Image processing algorithms for real time automatic throwing detection and dice classification were developed and implemented.

Automated measurement of retinal vascular tortuosity.

PubMed Central

Hart, W. E.; Goldbaum, M.; Côté, B.; Kube, P.; Nelson, M. R.

1997-01-01

Automatic measurement of blood vessel tortuosity is a useful capability for automatic ophthalmological diagnostic tools. We describe a suite of automated tortuosity measures for blood vessel segments extracted from RGB retinal images. The tortuosity measures were evaluated in two classification tasks: (1) classifying the tortuosity of blood vessel segments and (2) classifying the tortuosity of blood vessel networks. These tortuosity measures were able to achieve a classification rate of 91% for the first problem and 95% on the second problem, which confirms that they capture much of the ophthalmologists' notion of tortuosity. Images Figure 1 PMID:9357668

Imaging Depression in Adults with ASD

DTIC Science & Technology

2016-10-01

DISTRIBUTION UNLIMITED 13. SUPPLEMENTARY NOTES 14.ABSTRACT Aim A: To determine if the immunologic bias in autism spectrum disorder (ASD) confers greater...with no results to report to date. 1S. SUBJECT TERMS AUTISM , BRAIN IMAGING, DEPRESSION, SOCIAL REJECTION 16. SECURITY CLASSIFICATION OF: 17...which will be followed by an exploratory hedonic reward task, Monitory Incentive Delay. 2. KEYWORDS: autism , brain imaging, depression, social

Multiplex Quantitative Histologic Analysis of Human Breast Cancer Cell Signaling and Cell Fate

DTIC Science & Technology

2010-05-01

Breast cancer, cell signaling, cell proliferation, histology, image analysis 15. NUMBER OF PAGES - 51 16. PRICE CODE 17. SECURITY CLASSIFICATION...revealed by individual stains in multiplex combinations; and (3) software (FARSIGHT) for automated multispectral image analysis that (i) segments...Task 3. Develop computational algorithms for multispectral immunohistological image analysis FARSIGHT software was developed to quantify intrinsic

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.

Deep learning for brain tumor classification

NASA Astrophysics Data System (ADS)

Paul, Justin S.; Plassard, Andrew J.; Landman, Bennett A.; Fabbri, Daniel

2017-03-01

Recent research has shown that deep learning methods have performed well on supervised machine learning, image classification tasks. The purpose of this study is to apply deep learning methods to classify brain images with different tumor types: meningioma, glioma, and pituitary. A dataset was publicly released containing 3,064 T1-weighted contrast enhanced MRI (CE-MRI) brain images from 233 patients with either meningioma, glioma, or pituitary tumors split across axial, coronal, or sagittal planes. This research focuses on the 989 axial images from 191 patients in order to avoid confusing the neural networks with three different planes containing the same diagnosis. Two types of neural networks were used in classification: fully connected and convolutional neural networks. Within these two categories, further tests were computed via the augmentation of the original 512×512 axial images. Training neural networks over the axial data has proven to be accurate in its classifications with an average five-fold cross validation of 91.43% on the best trained neural network. This result demonstrates that a more general method (i.e. deep learning) can outperform specialized methods that require image dilation and ring-forming subregions on tumors.

Semi-Supervised Marginal Fisher Analysis for Hyperspectral Image Classification

NASA Astrophysics Data System (ADS)

Huang, H.; Liu, J.; Pan, Y.

2012-07-01

The problem of learning with both labeled and unlabeled examples arises frequently in Hyperspectral image (HSI) classification. While marginal Fisher analysis is a supervised method, which cannot be directly applied for Semi-supervised classification. In this paper, we proposed a novel method, called semi-supervised marginal Fisher analysis (SSMFA), to process HSI of natural scenes, which uses a combination of semi-supervised learning and manifold learning. In SSMFA, a new difference-based optimization objective function with unlabeled samples has been designed. SSMFA preserves the manifold structure of labeled and unlabeled samples in addition to separating labeled samples in different classes from each other. The semi-supervised method has an analytic form of the globally optimal solution, and it can be computed based on eigen decomposition. Classification experiments with a challenging HSI task demonstrate that this method outperforms current state-of-the-art HSI-classification methods.

Histopathological Image Classification using Discriminative Feature-oriented Dictionary Learning

PubMed Central

Vu, Tiep Huu; Mousavi, Hojjat Seyed; Monga, Vishal; Rao, Ganesh; Rao, UK Arvind

2016-01-01

In histopathological image analysis, feature extraction for classification is a challenging task due to the diversity of histology features suitable for each problem as well as presence of rich geometrical structures. In this paper, we propose an automatic feature discovery framework via learning class-specific dictionaries and present a low-complexity method for classification and disease grading in histopathology. Essentially, our Discriminative Feature-oriented Dictionary Learning (DFDL) method learns class-specific dictionaries such that under a sparsity constraint, the learned dictionaries allow representing a new image sample parsimoniously via the dictionary corresponding to the class identity of the sample. At the same time, the dictionary is designed to be poorly capable of representing samples from other classes. Experiments on three challenging real-world image databases: 1) histopathological images of intraductal breast lesions, 2) mammalian kidney, lung and spleen images provided by the Animal Diagnostics Lab (ADL) at Pennsylvania State University, and 3) brain tumor images from The Cancer Genome Atlas (TCGA) database, reveal the merits of our proposal over state-of-the-art alternatives. Moreover, we demonstrate that DFDL exhibits a more graceful decay in classification accuracy against the number of training images which is highly desirable in practice where generous training is often not available. PMID:26513781

Spectral-spatial hyperspectral image classification using super-pixel-based spatial pyramid representation

NASA Astrophysics Data System (ADS)

Fan, Jiayuan; Tan, Hui Li; Toomik, Maria; Lu, Shijian

2016-10-01

Spatial pyramid matching has demonstrated its power for image recognition task by pooling features from spatially increasingly fine sub-regions. Motivated by the concept of feature pooling at multiple pyramid levels, we propose a novel spectral-spatial hyperspectral image classification approach using superpixel-based spatial pyramid representation. This technique first generates multiple superpixel maps by decreasing the superpixel number gradually along with the increased spatial regions for labelled samples. By using every superpixel map, sparse representation of pixels within every spatial region is then computed through local max pooling. Finally, features learned from training samples are aggregated and trained by a support vector machine (SVM) classifier. The proposed spectral-spatial hyperspectral image classification technique has been evaluated on two public hyperspectral datasets, including the Indian Pines image containing 16 different agricultural scene categories with a 20m resolution acquired by AVIRIS and the University of Pavia image containing 9 land-use categories with a 1.3m spatial resolution acquired by the ROSIS-03 sensor. Experimental results show significantly improved performance compared with the state-of-the-art works. The major contributions of this proposed technique include (1) a new spectral-spatial classification approach to generate feature representation for hyperspectral image, (2) a complementary yet effective feature pooling approach, i.e. the superpixel-based spatial pyramid representation that is used for the spatial correlation study, (3) evaluation on two public hyperspectral image datasets with superior image classification performance.

Investigation of Latent Traces Using Infrared Reflectance Hyperspectral Imaging

NASA Astrophysics Data System (ADS)

Schubert, Till; Wenzel, Susanne; Roscher, Ribana; Stachniss, Cyrill

2016-06-01

The detection of traces is a main task of forensics. Hyperspectral imaging is a potential method from which we expect to capture more fluorescence effects than with common forensic light sources. This paper shows that the use of hyperspectral imaging is suited for the analysis of latent traces and extends the classical concept to the conservation of the crime scene for retrospective laboratory analysis. We examine specimen of blood, semen and saliva traces in several dilution steps, prepared on cardboard substrate. As our key result we successfully make latent traces visible up to dilution factor of 1:8000. We can attribute most of the detectability to interference of electromagnetic light with the water content of the traces in the shortwave infrared region of the spectrum. In a classification task we use several dimensionality reduction methods (PCA and LDA) in combination with a Maximum Likelihood classifier, assuming normally distributed data. Further, we use Random Forest as a competitive approach. The classifiers retrieve the exact positions of labelled trace preparation up to highest dilution and determine posterior probabilities. By modelling the classification task with a Markov Random Field we are able to integrate prior information about the spatial relation of neighboured pixel labels.

Deep learning based classification of breast tumors with shear-wave elastography.

PubMed

Zhang, Qi; Xiao, Yang; Dai, Wei; Suo, Jingfeng; Wang, Congzhi; Shi, Jun; Zheng, Hairong

2016-12-01

This study aims to build a deep learning (DL) architecture for automated extraction of learned-from-data image features from the shear-wave elastography (SWE), and to evaluate the DL architecture in differentiation between benign and malignant breast tumors. We construct a two-layer DL architecture for SWE feature extraction, comprised of the point-wise gated Boltzmann machine (PGBM) and the restricted Boltzmann machine (RBM). The PGBM contains task-relevant and task-irrelevant hidden units, and the task-relevant units are connected to the RBM. Experimental evaluation was performed with five-fold cross validation on a set of 227 SWE images, 135 of benign tumors and 92 of malignant tumors, from 121 patients. The features learned with our DL architecture were compared with the statistical features quantifying image intensity and texture. Results showed that the DL features achieved better classification performance with an accuracy of 93.4%, a sensitivity of 88.6%, a specificity of 97.1%, and an area under the receiver operating characteristic curve of 0.947. The DL-based method integrates feature learning with feature selection on SWE. It may be potentially used in clinical computer-aided diagnosis of breast cancer. Copyright © 2016 Elsevier B.V. All rights reserved.

Image processing and recognition for biological images

PubMed Central

Uchida, Seiichi

2013-01-01

This paper reviews image processing and pattern recognition techniques, which will be useful to analyze bioimages. Although this paper does not provide their technical details, it will be possible to grasp their main tasks and typical tools to handle the tasks. Image processing is a large research area to improve the visibility of an input image and acquire some valuable information from it. As the main tasks of image processing, this paper introduces gray-level transformation, binarization, image filtering, image segmentation, visual object tracking, optical flow and image registration. Image pattern recognition is the technique to classify an input image into one of the predefined classes and also has a large research area. This paper overviews its two main modules, that is, feature extraction module and classification module. Throughout the paper, it will be emphasized that bioimage is a very difficult target for even state-of-the-art image processing and pattern recognition techniques due to noises, deformations, etc. This paper is expected to be one tutorial guide to bridge biology and image processing researchers for their further collaboration to tackle such a difficult target. PMID:23560739

A method to classify schizophrenia using inter-task spatial correlations of functional brain images.

PubMed

Michael, Andrew M; Calhoun, Vince D; Andreasen, Nancy C; Baum, Stefi A

2008-01-01

The clinical heterogeneity of schizophrenia (scz) and the overlap of self reported and observed symptoms with other mental disorders makes its diagnosis a difficult task. At present no laboratory-based or image-based diagnostic tool for scz exists and such tools are desired to support existing methods for more precise diagnosis. Functional magnetic resonance imaging (fMRI) is currently employed to identify and correlate cognitive processes related to scz and its symptoms. Fusion of multiple fMRI tasks that probe different cognitive processes may help to better understand hidden networks of this complex disorder. In this paper we utilize three different fMRI tasks and introduce an approach to classify subjects based on inter-task spatial correlations of brain activation. The technique was applied to groups of patients and controls and its validity was checked with the leave-one-out method. We show that the classification rate increases when information from multiple tasks are combined.

Functional Magnetic Resonance Imaging of Cognitive Processing in Young Adults with Down Syndrome

ERIC Educational Resources Information Center

Jacola, Lisa M.; Byars, Anna W.; Chalfonte-Evans, Melinda; Schmithorst, Vincent J.; Hickey, Fran; Patterson, Bonnie; Hotze, Stephanie; Vannest, Jennifer; Chiu, Chung-Yiu; Holland, Scott K.; Schapiro, Mark B.

2011-01-01

The authors used functional magnetic resonance imaging (fMRI) to investigate neural activation during a semantic-classification/object-recognition task in 13 persons with Down syndrome and 12 typically developing control participants (age range = 12-26 years). A comparison between groups suggested atypical patterns of brain activation for the…

Assessing the performance of multiple spectral-spatial features of a hyperspectral image for classification of urban land cover classes using support vector machines and artificial neural network

NASA Astrophysics Data System (ADS)

Pullanagari, Reddy; Kereszturi, Gábor; Yule, Ian J.; Ghamisi, Pedram

2017-04-01

Accurate and spatially detailed mapping of complex urban environments is essential for land managers. Classifying high spectral and spatial resolution hyperspectral images is a challenging task because of its data abundance and computational complexity. Approaches with a combination of spectral and spatial information in a single classification framework have attracted special attention because of their potential to improve the classification accuracy. We extracted multiple features from spectral and spatial domains of hyperspectral images and evaluated them with two supervised classification algorithms; support vector machines (SVM) and an artificial neural network. The spatial features considered are produced by a gray level co-occurrence matrix and extended multiattribute profiles. All of these features were stacked, and the most informative features were selected using a genetic algorithm-based SVM. After selecting the most informative features, the classification model was integrated with a segmentation map derived using a hidden Markov random field. We tested the proposed method on a real application of a hyperspectral image acquired from AisaFENIX and on widely used hyperspectral images. From the results, it can be concluded that the proposed framework significantly improves the results with different spectral and spatial resolutions over different instrumentation.

SAR target recognition and posture estimation using spatial pyramid pooling within CNN

NASA Astrophysics Data System (ADS)

Peng, Lijiang; Liu, Xiaohua; Liu, Ming; Dong, Liquan; Hui, Mei; Zhao, Yuejin

2018-01-01

Many convolution neural networks(CNN) architectures have been proposed to strengthen the performance on synthetic aperture radar automatic target recognition (SAR-ATR) and obtained state-of-art results on targets classification on MSTAR database, but few methods concern about the estimation of depression angle and azimuth angle of targets. To get better effect on learning representation of hierarchies of features on both 10-class target classification task and target posture estimation tasks, we propose a new CNN architecture with spatial pyramid pooling(SPP) which can build high hierarchy of features map by dividing the convolved feature maps from finer to coarser levels to aggregate local features of SAR images. Experimental results on MSTAR database show that the proposed architecture can get high recognition accuracy as 99.57% on 10-class target classification task as the most current state-of-art methods, and also get excellent performance on target posture estimation tasks which pays attention to depression angle variety and azimuth angle variety. What's more, the results inspire us the application of deep learning on SAR target posture description.

Galaxy Classifications with Deep Learning

NASA Astrophysics Data System (ADS)

Lukic, Vesna; Brüggen, Marcus

2017-06-01

Machine learning techniques have proven to be increasingly useful in astronomical applications over the last few years, for example in object classification, estimating redshifts and data mining. One example of object classification is classifying galaxy morphology. This is a tedious task to do manually, especially as the datasets become larger with surveys that have a broader and deeper search-space. The Kaggle Galaxy Zoo competition presented the challenge of writing an algorithm to find the probability that a galaxy belongs in a particular class, based on SDSS optical spectroscopy data. The use of convolutional neural networks (convnets), proved to be a popular solution to the problem, as they have also produced unprecedented classification accuracies in other image databases such as the database of handwritten digits (MNIST †) and large database of images (CIFAR ‡). We experiment with the convnets that comprised the winning solution, but using broad classifications. The effect of changing the number of layers is explored, as well as using a different activation function, to help in developing an intuition of how the networks function and to see how they can be applied to radio galaxy images.

Neural Signatures of Trust During Human-Automation Interactions

DTIC Science & Technology

2016-04-01

magnetic resonance imaging by manipulating the reliability of advice from a human or automated luggage inspector framed as experts. HAT and HHT were...human-human trust, human-automation trust, brain, functional magnetic resonance imaging 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT 18...behavioral X-ray luggage-screening task with functional magnetic resonance imaging (fMRI) and manipulated reliabilities of advice (unknown to the

Wavelength-adaptive dehazing using histogram merging-based classification for UAV images.

PubMed

Yoon, Inhye; Jeong, Seokhwa; Jeong, Jaeheon; Seo, Doochun; Paik, Joonki

2015-03-19

Since incoming light to an unmanned aerial vehicle (UAV) platform can be scattered by haze and dust in the atmosphere, the acquired image loses the original color and brightness of the subject. Enhancement of hazy images is an important task in improving the visibility of various UAV images. This paper presents a spatially-adaptive dehazing algorithm that merges color histograms with consideration of the wavelength-dependent atmospheric turbidity. Based on the wavelength-adaptive hazy image acquisition model, the proposed dehazing algorithm consists of three steps: (i) image segmentation based on geometric classes; (ii) generation of the context-adaptive transmission map; and (iii) intensity transformation for enhancing a hazy UAV image. The major contribution of the research is a novel hazy UAV image degradation model by considering the wavelength of light sources. In addition, the proposed transmission map provides a theoretical basis to differentiate visually important regions from others based on the turbidity and merged classification results.

Convolutional Neural Network for Histopathological Analysis of Osteosarcoma.

PubMed

Mishra, Rashika; Daescu, Ovidiu; Leavey, Patrick; Rakheja, Dinesh; Sengupta, Anita

2018-03-01

Pathologists often deal with high complexity and sometimes disagreement over osteosarcoma tumor classification due to cellular heterogeneity in the dataset. Segmentation and classification of histology tissue in H&E stained tumor image datasets is a challenging task because of intra-class variations, inter-class similarity, crowded context, and noisy data. In recent years, deep learning approaches have led to encouraging results in breast cancer and prostate cancer analysis. In this article, we propose convolutional neural network (CNN) as a tool to improve efficiency and accuracy of osteosarcoma tumor classification into tumor classes (viable tumor, necrosis) versus nontumor. The proposed CNN architecture contains eight learned layers: three sets of stacked two convolutional layers interspersed with max pooling layers for feature extraction and two fully connected layers with data augmentation strategies to boost performance. The use of a neural network results in higher accuracy of average 92% for the classification. We compare the proposed architecture with three existing and proven CNN architectures for image classification: AlexNet, LeNet, and VGGNet. We also provide a pipeline to calculate percentage necrosis in a given whole slide image. We conclude that the use of neural networks can assure both high accuracy and efficiency in osteosarcoma classification.

Vehicle Surveillance with a Generic, Adaptive, 3D Vehicle Model.

PubMed

Leotta, Matthew J; Mundy, Joseph L

2011-07-01

In automated surveillance, one is often interested in tracking road vehicles, measuring their shape in 3D world space, and determining vehicle classification. To address these tasks simultaneously, an effective approach is the constrained alignment of a prior model of 3D vehicle shape to images. Previous 3D vehicle models are either generic but overly simple or rigid and overly complex. Rigid models represent exactly one vehicle design, so a large collection is needed. A single generic model can deform to a wide variety of shapes, but those shapes have been far too primitive. This paper uses a generic 3D vehicle model that deforms to match a wide variety of passenger vehicles. It is adjustable in complexity between the two extremes. The model is aligned to images by predicting and matching image intensity edges. Novel algorithms are presented for fitting models to multiple still images and simultaneous tracking while estimating shape in video. Experiments compare the proposed model to simple generic models in accuracy and reliability of 3D shape recovery from images and tracking in video. Standard techniques for classification are also used to compare the models. The proposed model outperforms the existing simple models at each task.

Deep Learning for Classification of Colorectal Polyps on Whole-slide Images.

PubMed

Korbar, Bruno; Olofson, Andrea M; Miraflor, Allen P; Nicka, Catherine M; Suriawinata, Matthew A; Torresani, Lorenzo; Suriawinata, Arief A; Hassanpour, Saeed

2017-01-01

Histopathological characterization of colorectal polyps is critical for determining the risk of colorectal cancer and future rates of surveillance for patients. However, this characterization is a challenging task and suffers from significant inter- and intra-observer variability. We built an automatic image analysis method that can accurately classify different types of colorectal polyps on whole-slide images to help pathologists with this characterization and diagnosis. Our method is based on deep-learning techniques, which rely on numerous levels of abstraction for data representation and have shown state-of-the-art results for various image analysis tasks. Our method covers five common types of polyps (i.e., hyperplastic, sessile serrated, traditional serrated, tubular, and tubulovillous/villous) that are included in the US Multisociety Task Force guidelines for colorectal cancer risk assessment and surveillance. We developed multiple deep-learning approaches by leveraging a dataset of 2074 crop images, which were annotated by multiple domain expert pathologists as reference standards. We evaluated our method on an independent test set of 239 whole-slide images and measured standard machine-learning evaluation metrics of accuracy, precision, recall, and F1 score and their 95% confidence intervals. Our evaluation shows that our method with residual network architecture achieves the best performance for classification of colorectal polyps on whole-slide images (overall accuracy: 93.0%, 95% confidence interval: 89.0%-95.9%). Our method can reduce the cognitive burden on pathologists and improve their efficacy in histopathological characterization of colorectal polyps and in subsequent risk assessment and follow-up recommendations.

Deep Learning for Classification of Colorectal Polyps on Whole-slide Images

PubMed Central

Korbar, Bruno; Olofson, Andrea M.; Miraflor, Allen P.; Nicka, Catherine M.; Suriawinata, Matthew A.; Torresani, Lorenzo; Suriawinata, Arief A.; Hassanpour, Saeed

2017-01-01

Context: Histopathological characterization of colorectal polyps is critical for determining the risk of colorectal cancer and future rates of surveillance for patients. However, this characterization is a challenging task and suffers from significant inter- and intra-observer variability. Aims: We built an automatic image analysis method that can accurately classify different types of colorectal polyps on whole-slide images to help pathologists with this characterization and diagnosis. Setting and Design: Our method is based on deep-learning techniques, which rely on numerous levels of abstraction for data representation and have shown state-of-the-art results for various image analysis tasks. Subjects and Methods: Our method covers five common types of polyps (i.e., hyperplastic, sessile serrated, traditional serrated, tubular, and tubulovillous/villous) that are included in the US Multisociety Task Force guidelines for colorectal cancer risk assessment and surveillance. We developed multiple deep-learning approaches by leveraging a dataset of 2074 crop images, which were annotated by multiple domain expert pathologists as reference standards. Statistical Analysis: We evaluated our method on an independent test set of 239 whole-slide images and measured standard machine-learning evaluation metrics of accuracy, precision, recall, and F1 score and their 95% confidence intervals. Results: Our evaluation shows that our method with residual network architecture achieves the best performance for classification of colorectal polyps on whole-slide images (overall accuracy: 93.0%, 95% confidence interval: 89.0%–95.9%). Conclusions: Our method can reduce the cognitive burden on pathologists and improve their efficacy in histopathological characterization of colorectal polyps and in subsequent risk assessment and follow-up recommendations. PMID:28828201

Decomposition-based transfer distance metric learning for image classification.

PubMed

Luo, Yong; Liu, Tongliang; Tao, Dacheng; Xu, Chao

2014-09-01

Distance metric learning (DML) is a critical factor for image analysis and pattern recognition. To learn a robust distance metric for a target task, we need abundant side information (i.e., the similarity/dissimilarity pairwise constraints over the labeled data), which is usually unavailable in practice due to the high labeling cost. This paper considers the transfer learning setting by exploiting the large quantity of side information from certain related, but different source tasks to help with target metric learning (with only a little side information). The state-of-the-art metric learning algorithms usually fail in this setting because the data distributions of the source task and target task are often quite different. We address this problem by assuming that the target distance metric lies in the space spanned by the eigenvectors of the source metrics (or other randomly generated bases). The target metric is represented as a combination of the base metrics, which are computed using the decomposed components of the source metrics (or simply a set of random bases); we call the proposed method, decomposition-based transfer DML (DTDML). In particular, DTDML learns a sparse combination of the base metrics to construct the target metric by forcing the target metric to be close to an integration of the source metrics. The main advantage of the proposed method compared with existing transfer metric learning approaches is that we directly learn the base metric coefficients instead of the target metric. To this end, far fewer variables need to be learned. We therefore obtain more reliable solutions given the limited side information and the optimization tends to be faster. Experiments on the popular handwritten image (digit, letter) classification and challenge natural image annotation tasks demonstrate the effectiveness of the proposed method.

Using machine learning techniques to automate sky survey catalog generation

NASA Technical Reports Server (NTRS)

Fayyad, Usama M.; Roden, J. C.; Doyle, R. J.; Weir, Nicholas; Djorgovski, S. G.

1993-01-01

We describe the application of machine classification techniques to the development of an automated tool for the reduction of a large scientific data set. The 2nd Palomar Observatory Sky Survey provides comprehensive photographic coverage of the northern celestial hemisphere. The photographic plates are being digitized into images containing on the order of 10(exp 7) galaxies and 10(exp 8) stars. Since the size of this data set precludes manual analysis and classification of objects, our approach is to develop a software system which integrates independently developed techniques for image processing and data classification. Image processing routines are applied to identify and measure features of sky objects. Selected features are used to determine the classification of each object. GID3* and O-BTree, two inductive learning techniques, are used to automatically learn classification decision trees from examples. We describe the techniques used, the details of our specific application, and the initial encouraging results which indicate that our approach is well-suited to the problem. The benefits of the approach are increased data reduction throughput, consistency of classification, and the automated derivation of classification rules that will form an objective, examinable basis for classifying sky objects. Furthermore, astronomers will be freed from the tedium of an intensely visual task to pursue more challenging analysis and interpretation problems given automatically cataloged data.

Simultaneous acquisition of EEG and NIRS during cognitive tasks for an open access dataset.

PubMed

Shin, Jaeyoung; von Lühmann, Alexander; Kim, Do-Won; Mehnert, Jan; Hwang, Han-Jeong; Müller, Klaus-Robert

2018-02-13

We provide an open access multimodal brain-imaging dataset of simultaneous electroencephalography (EEG) and near-infrared spectroscopy (NIRS) recordings. Twenty-six healthy participants performed three cognitive tasks: 1) n-back (0-, 2- and 3-back), 2) discrimination/selection response task (DSR) and 3) word generation (WG) tasks. The data provided includes: 1) measured data, 2) demographic data, and 3) basic analysis results. For n-back (dataset A) and DSR tasks (dataset B), event-related potential (ERP) analysis was performed, and spatiotemporal characteristics and classification results for 'target' versus 'non-target' (dataset A) and symbol 'O' versus symbol 'X' (dataset B) are provided. Time-frequency analysis was performed to show the EEG spectral power to differentiate the task-relevant activations. Spatiotemporal characteristics of hemodynamic responses are also shown. For the WG task (dataset C), the EEG spectral power and spatiotemporal characteristics of hemodynamic responses are analyzed, and the potential merit of hybrid EEG-NIRS BCIs was validated with respect to classification accuracy. We expect that the dataset provided will facilitate performance evaluation and comparison of many neuroimaging analysis techniques.

Simultaneous acquisition of EEG and NIRS during cognitive tasks for an open access dataset

PubMed Central

Shin, Jaeyoung; von Lühmann, Alexander; Kim, Do-Won; Mehnert, Jan; Hwang, Han-Jeong; Müller, Klaus-Robert

2018-01-01

We provide an open access multimodal brain-imaging dataset of simultaneous electroencephalography (EEG) and near-infrared spectroscopy (NIRS) recordings. Twenty-six healthy participants performed three cognitive tasks: 1) n-back (0-, 2- and 3-back), 2) discrimination/selection response task (DSR) and 3) word generation (WG) tasks. The data provided includes: 1) measured data, 2) demographic data, and 3) basic analysis results. For n-back (dataset A) and DSR tasks (dataset B), event-related potential (ERP) analysis was performed, and spatiotemporal characteristics and classification results for ‘target’ versus ‘non-target’ (dataset A) and symbol ‘O’ versus symbol ‘X’ (dataset B) are provided. Time-frequency analysis was performed to show the EEG spectral power to differentiate the task-relevant activations. Spatiotemporal characteristics of hemodynamic responses are also shown. For the WG task (dataset C), the EEG spectral power and spatiotemporal characteristics of hemodynamic responses are analyzed, and the potential merit of hybrid EEG-NIRS BCIs was validated with respect to classification accuracy. We expect that the dataset provided will facilitate performance evaluation and comparison of many neuroimaging analysis techniques. PMID:29437166

Pattern recognition for passive polarimetric data using nonparametric classifiers

NASA Astrophysics Data System (ADS)

Thilak, Vimal; Saini, Jatinder; Voelz, David G.; Creusere, Charles D.

2005-08-01

Passive polarization based imaging is a useful tool in computer vision and pattern recognition. A passive polarization imaging system forms a polarimetric image from the reflection of ambient light that contains useful information for computer vision tasks such as object detection (classification) and recognition. Applications of polarization based pattern recognition include material classification and automatic shape recognition. In this paper, we present two target detection algorithms for images captured by a passive polarimetric imaging system. The proposed detection algorithms are based on Bayesian decision theory. In these approaches, an object can belong to one of any given number classes and classification involves making decisions that minimize the average probability of making incorrect decisions. This minimum is achieved by assigning an object to the class that maximizes the a posteriori probability. Computing a posteriori probabilities requires estimates of class conditional probability density functions (likelihoods) and prior probabilities. A Probabilistic neural network (PNN), which is a nonparametric method that can compute Bayes optimal boundaries, and a -nearest neighbor (KNN) classifier, is used for density estimation and classification. The proposed algorithms are applied to polarimetric image data gathered in the laboratory with a liquid crystal-based system. The experimental results validate the effectiveness of the above algorithms for target detection from polarimetric data.

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.

Segmentation of radiologic images with self-organizing maps: the segmentation problem transformed into a classification task

NASA Astrophysics Data System (ADS)

Pelikan, Erich; Vogelsang, Frank; Tolxdorff, Thomas

1996-04-01

The texture-based segmentation of x-ray images of focal bone lesions using topological maps is introduced. Texture characteristics are described by image-point correlation of feature images to feature vectors. For the segmentation, the topological map is labeled using an improved labeling strategy. Results of the technique are demonstrated on original and synthetic x-ray images and quantified with the aid of quality measures. In addition, a classifier-specific contribution analysis is applied for assessing the feature space.

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).

Image processing pipeline for segmentation and material classification based on multispectral high dynamic range polarimetric images.

PubMed

Martínez-Domingo, Miguel Ángel; Valero, Eva M; Hernández-Andrés, Javier; Tominaga, Shoji; Horiuchi, Takahiko; Hirai, Keita

2017-11-27

We propose a method for the capture of high dynamic range (HDR), multispectral (MS), polarimetric (Pol) images of indoor scenes using a liquid crystal tunable filter (LCTF). We have included the adaptive exposure estimation (AEE) method to fully automatize the capturing process. We also propose a pre-processing method which can be applied for the registration of HDR images after they are already built as the result of combining different low dynamic range (LDR) images. This method is applied to ensure a correct alignment of the different polarization HDR images for each spectral band. We have focused our efforts in two main applications: object segmentation and classification into metal and dielectric classes. We have simplified the segmentation using mean shift combined with cluster averaging and region merging techniques. We compare the performance of our segmentation with that of Ncut and Watershed methods. For the classification task, we propose to use information not only in the highlight regions but also in their surrounding area, extracted from the degree of linear polarization (DoLP) maps. We present experimental results which proof that the proposed image processing pipeline outperforms previous techniques developed specifically for MSHDRPol image cubes.

Template optimization and transfer in perceptual learning.

PubMed

Kurki, Ilmari; Hyvärinen, Aapo; Saarinen, Jussi

2016-08-01

We studied how learning changes the processing of a low-level Gabor stimulus, using a classification-image method (psychophysical reverse correlation) and a task where observers discriminated between slight differences in the phase (relative alignment) of a target Gabor in visual noise. The method estimates the internal "template" that describes how the visual system weights the input information for decisions. One popular idea has been that learning makes the template more like an ideal Bayesian weighting; however, the evidence has been indirect. We used a new regression technique to directly estimate the template weight change and to test whether the direction of reweighting is significantly different from an optimal learning strategy. The subjects trained the task for six daily sessions, and we tested the transfer of training to a target in an orthogonal orientation. Strong learning and partial transfer were observed. We tested whether task precision (difficulty) had an effect on template change and transfer: Observers trained in either a high-precision (small, 60° phase difference) or a low-precision task (180°). Task precision did not have an effect on the amount of template change or transfer, suggesting that task precision per se does not determine whether learning generalizes. Classification images show that training made observers use more task-relevant features and unlearn some irrelevant features. The transfer templates resembled partially optimized versions of templates in training sessions. The template change direction resembles ideal learning significantly but not completely. The amount of template change was highly correlated with the amount of learning.

Resting-State Functional Magnetic Resonance Imaging for Language Preoperative Planning

PubMed Central

Branco, Paulo; Seixas, Daniela; Deprez, Sabine; Kovacs, Silvia; Peeters, Ronald; Castro, São L.; Sunaert, Stefan

2016-01-01

Functional magnetic resonance imaging (fMRI) is a well-known non-invasive technique for the study of brain function. One of its most common clinical applications is preoperative language mapping, essential for the preservation of function in neurosurgical patients. Typically, fMRI is used to track task-related activity, but poor task performance and movement artifacts can be critical limitations in clinical settings. Recent advances in resting-state protocols open new possibilities for pre-surgical mapping of language potentially overcoming these limitations. To test the feasibility of using resting-state fMRI instead of conventional active task-based protocols, we compared results from fifteen patients with brain lesions while performing a verb-to-noun generation task and while at rest. Task-activity was measured using a general linear model analysis and independent component analysis (ICA). Resting-state networks were extracted using ICA and further classified in two ways: manually by an expert and by using an automated template matching procedure. The results revealed that the automated classification procedure correctly identified language networks as compared to the expert manual classification. We found a good overlay between task-related activity and resting-state language maps, particularly within the language regions of interest. Furthermore, resting-state language maps were as sensitive as task-related maps, and had higher specificity. Our findings suggest that resting-state protocols may be suitable to map language networks in a quick and clinically efficient way. PMID:26869899

The use of global image characteristics for neural network pattern recognitions

NASA Astrophysics Data System (ADS)

Kulyas, Maksim O.; Kulyas, Oleg L.; Loshkarev, Aleksey S.

2017-04-01

The recognition system is observed, where the information is transferred by images of symbols generated by a television camera. For descriptors of objects the coefficients of two-dimensional Fourier transformation generated in a special way. For solution of the task of classification the one-layer neural network trained on reference images is used. Fast learning of a neural network with a single neuron calculation of coefficients is applied.

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

Drawing a baseline in aesthetic quality assessment

NASA Astrophysics Data System (ADS)

Rubio, Fernando; Flores, M. Julia; Puerta, Jose M.

2018-04-01

Aesthetic classification of images is an inherently subjective task. There does not exist a validated collection of images/photographs labeled as having good or bad quality from experts. Nowadays, the closest approximation to that is to use databases of photos where a group of users rate each image. Hence, there is not a unique good/bad label but a rating distribution given by users voting. Due to this peculiarity, it is not possible to state the problem of binary aesthetic supervised classification in such a direct mode as other Computer Vision tasks. Recent literature follows an approach where researchers utilize the average rates from the users for each image, and they establish an arbitrary threshold to determine their class or label. In this way, images above the threshold are considered of good quality, while images below the threshold are seen as bad quality. This paper analyzes current literature, and it reviews those attributes able to represent an image, differentiating into three families: specific, general and deep features. Among those which have been proved more competitive, we have selected a representative subset, being our main goal to establish a clear experimental framework. Finally, once features were selected, we have used them for the full AVA dataset. We have to remark that to perform validation we report not only accuracy values, which is not that informative in this case, but also, metrics able to evaluate classification power within imbalanced datasets. We have conducted a series of experiments so that distinct well-known classifiers are learned from data. Like that, this paper provides what we could consider valuable and valid baseline results for the given problem.

Manifold Regularized Multitask Feature Learning for Multimodality Disease Classification

PubMed Central

Jie, Biao; Zhang, Daoqiang; Cheng, Bo; Shen, Dinggang

2015-01-01

Multimodality based methods have shown great advantages in classification of Alzheimer’s disease (AD) and its prodromal stage, that is, mild cognitive impairment (MCI). Recently, multitask feature selection methods are typically used for joint selection of common features across multiple modalities. However, one disadvantage of existing multimodality based methods is that they ignore the useful data distribution information in each modality, which is essential for subsequent classification. Accordingly, in this paper we propose a manifold regularized multitask feature learning method to preserve both the intrinsic relatedness among multiple modalities of data and the data distribution information in each modality. Specifically, we denote the feature learning on each modality as a single task, and use group-sparsity regularizer to capture the intrinsic relatedness among multiple tasks (i.e., modalities) and jointly select the common features from multiple tasks. Furthermore, we introduce a new manifold-based Laplacian regularizer to preserve the data distribution information from each task. Finally, we use the multikernel support vector machine method to fuse multimodality data for eventual classification. Conversely, we also extend our method to the semisupervised setting, where only partial data are labeled. We evaluate our method using the baseline magnetic resonance imaging (MRI), fluorodeoxyglucose positron emission tomography (FDG-PET), and cerebrospinal fluid (CSF) data of subjects from AD neuroimaging initiative database. The experimental results demonstrate that our proposed method can not only achieve improved classification performance, but also help to discover the disease-related brain regions useful for disease diagnosis. PMID:25277605

A new classification scheme of plastic wastes based upon recycling labels

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

Özkan, Kemal, E-mail: kozkan@ogu.edu.tr; Ergin, Semih, E-mail: sergin@ogu.edu.tr; Işık, Şahin, E-mail: sahini@ogu.edu.tr

Highlights: • PET, HPDE or PP types of plastics are considered. • An automated classification of plastic bottles based on the feature extraction and classification methods is performed. • The decision mechanism consists of PCA, Kernel PCA, FLDA, SVD and Laplacian Eigenmaps methods. • SVM is selected to achieve the classification task and majority voting technique is used. - Abstract: Since recycling of materials is widely assumed to be environmentally and economically beneficial, reliable sorting and processing of waste packaging materials such as plastics is very important for recycling with high efficiency. An automated system that can quickly categorize thesemore » materials is certainly needed for obtaining maximum classification while maintaining high throughput. In this paper, first of all, the photographs of the plastic bottles have been taken and several preprocessing steps were carried out. The first preprocessing step is to extract the plastic area of a bottle from the background. Then, the morphological image operations are implemented. These operations are edge detection, noise removal, hole removing, image enhancement, and image segmentation. These morphological operations can be generally defined in terms of the combinations of erosion and dilation. The effect of bottle color as well as label are eliminated using these operations. Secondly, the pixel-wise intensity values of the plastic bottle images have been used together with the most popular subspace and statistical feature extraction methods to construct the feature vectors in this study. Only three types of plastics are considered due to higher existence ratio of them than the other plastic types in the world. The decision mechanism consists of five different feature extraction methods including as Principal Component Analysis (PCA), Kernel PCA (KPCA), Fisher’s Linear Discriminant Analysis (FLDA), Singular Value Decomposition (SVD) and Laplacian Eigenmaps (LEMAP) and uses a simple experimental setup with a camera and homogenous backlighting. Due to the giving global solution for a classification problem, Support Vector Machine (SVM) is selected to achieve the classification task and majority voting technique is used as the decision mechanism. This technique equally weights each classification result and assigns the given plastic object to the class that the most classification results agree on. The proposed classification scheme provides high accuracy rate, and also it is able to run in real-time applications. It can automatically classify the plastic bottle types with approximately 90% recognition accuracy. Besides this, the proposed methodology yields approximately 96% classification rate for the separation of PET or non-PET plastic types. It also gives 92% accuracy for the categorization of non-PET plastic types into HPDE or PP.« less

Neural and Behavioral Sequelae of Blast-Related Traumatic Brain Injury

DTIC Science & Technology

2012-09-01

fMRI, DTI , cognition 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT 18. NUMBER OF PAGES 19a. NAME OF RESPONSIBLE PERSON USAMRMC a...techniques [task-activated functional MRI (fMRI) and diffusion tensor imaging ( DTI )] to gain a comprehensive understanding of the neural changes...orthopedic injuries. We accomplished this goal by conducting advanced neuroimaging (task-activated fMRI and DTI fiber tracking) and neurobehavioral

Image processing and recognition for biological images.

PubMed

Uchida, Seiichi

2013-05-01

This paper reviews image processing and pattern recognition techniques, which will be useful to analyze bioimages. Although this paper does not provide their technical details, it will be possible to grasp their main tasks and typical tools to handle the tasks. Image processing is a large research area to improve the visibility of an input image and acquire some valuable information from it. As the main tasks of image processing, this paper introduces gray-level transformation, binarization, image filtering, image segmentation, visual object tracking, optical flow and image registration. Image pattern recognition is the technique to classify an input image into one of the predefined classes and also has a large research area. This paper overviews its two main modules, that is, feature extraction module and classification module. Throughout the paper, it will be emphasized that bioimage is a very difficult target for even state-of-the-art image processing and pattern recognition techniques due to noises, deformations, etc. This paper is expected to be one tutorial guide to bridge biology and image processing researchers for their further collaboration to tackle such a difficult target. © 2013 The Author Development, Growth & Differentiation © 2013 Japanese Society of Developmental Biologists.

Cloud-Based NoSQL Open Database of Pulmonary Nodules for Computer-Aided Lung Cancer Diagnosis and Reproducible Research.

PubMed

Ferreira Junior, José Raniery; Oliveira, Marcelo Costa; de Azevedo-Marques, Paulo Mazzoncini

2016-12-01

Lung cancer is the leading cause of cancer-related deaths in the world, and its main manifestation is pulmonary nodules. Detection and classification of pulmonary nodules are challenging tasks that must be done by qualified specialists, but image interpretation errors make those tasks difficult. In order to aid radiologists on those hard tasks, it is important to integrate the computer-based tools with the lesion detection, pathology diagnosis, and image interpretation processes. However, computer-aided diagnosis research faces the problem of not having enough shared medical reference data for the development, testing, and evaluation of computational methods for diagnosis. In order to minimize this problem, this paper presents a public nonrelational document-oriented cloud-based database of pulmonary nodules characterized by 3D texture attributes, identified by experienced radiologists and classified in nine different subjective characteristics by the same specialists. Our goal with the development of this database is to improve computer-aided lung cancer diagnosis and pulmonary nodule detection and classification research through the deployment of this database in a cloud Database as a Service framework. Pulmonary nodule data was provided by the Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI), image descriptors were acquired by a volumetric texture analysis, and database schema was developed using a document-oriented Not only Structured Query Language (NoSQL) approach. The proposed database is now with 379 exams, 838 nodules, and 8237 images, 4029 of them are CT scans and 4208 manually segmented nodules, and it is allocated in a MongoDB instance on a cloud infrastructure.

PCANet: A Simple Deep Learning Baseline for Image Classification?

PubMed

Chan, Tsung-Han; Jia, Kui; Gao, Shenghua; Lu, Jiwen; Zeng, Zinan; Ma, Yi

2015-12-01

In this paper, we propose a very simple deep learning network for image classification that is based on very basic data processing components: 1) cascaded principal component analysis (PCA); 2) binary hashing; and 3) blockwise histograms. In the proposed architecture, the PCA is employed to learn multistage filter banks. This is followed by simple binary hashing and block histograms for indexing and pooling. This architecture is thus called the PCA network (PCANet) and can be extremely easily and efficiently designed and learned. For comparison and to provide a better understanding, we also introduce and study two simple variations of PCANet: 1) RandNet and 2) LDANet. They share the same topology as PCANet, but their cascaded filters are either randomly selected or learned from linear discriminant analysis. We have extensively tested these basic networks on many benchmark visual data sets for different tasks, including Labeled Faces in the Wild (LFW) for face verification; the MultiPIE, Extended Yale B, AR, Facial Recognition Technology (FERET) data sets for face recognition; and MNIST for hand-written digit recognition. Surprisingly, for all tasks, such a seemingly naive PCANet model is on par with the state-of-the-art features either prefixed, highly hand-crafted, or carefully learned [by deep neural networks (DNNs)]. Even more surprisingly, the model sets new records for many classification tasks on the Extended Yale B, AR, and FERET data sets and on MNIST variations. Additional experiments on other public data sets also demonstrate the potential of PCANet to serve as a simple but highly competitive baseline for texture classification and object recognition.

Automated Grading of Gliomas using Deep Learning in Digital Pathology Images: A modular approach with ensemble of convolutional neural networks.

PubMed

Ertosun, Mehmet Günhan; Rubin, Daniel L

2015-01-01

Brain glioma is the most common primary malignant brain tumors in adults with different pathologic subtypes: Lower Grade Glioma (LGG) Grade II, Lower Grade Glioma (LGG) Grade III, and Glioblastoma Multiforme (GBM) Grade IV. The survival and treatment options are highly dependent of this glioma grade. We propose a deep learning-based, modular classification pipeline for automated grading of gliomas using digital pathology images. Whole tissue digitized images of pathology slides obtained from The Cancer Genome Atlas (TCGA) were used to train our deep learning modules. Our modular pipeline provides diagnostic quality statistics, such as precision, sensitivity and specificity, of the individual deep learning modules, and (1) facilitates training given the limited data in this domain, (2) enables exploration of different deep learning structures for each module, (3) leads to developing less complex modules that are simpler to analyze, and (4) provides flexibility, permitting use of single modules within the framework or use of other modeling or machine learning applications, such as probabilistic graphical models or support vector machines. Our modular approach helps us meet the requirements of minimum accuracy levels that are demanded by the context of different decision points within a multi-class classification scheme. Convolutional Neural Networks are trained for each module for each sub-task with more than 90% classification accuracies on validation data set, and achieved classification accuracy of 96% for the task of GBM vs LGG classification, 71% for further identifying the grade of LGG into Grade II or Grade III on independent data set coming from new patients from the multi-institutional repository.

Automated Grading of Gliomas using Deep Learning in Digital Pathology Images: A modular approach with ensemble of convolutional neural networks

PubMed Central

Ertosun, Mehmet Günhan; Rubin, Daniel L.

2015-01-01

Brain glioma is the most common primary malignant brain tumors in adults with different pathologic subtypes: Lower Grade Glioma (LGG) Grade II, Lower Grade Glioma (LGG) Grade III, and Glioblastoma Multiforme (GBM) Grade IV. The survival and treatment options are highly dependent of this glioma grade. We propose a deep learning-based, modular classification pipeline for automated grading of gliomas using digital pathology images. Whole tissue digitized images of pathology slides obtained from The Cancer Genome Atlas (TCGA) were used to train our deep learning modules. Our modular pipeline provides diagnostic quality statistics, such as precision, sensitivity and specificity, of the individual deep learning modules, and (1) facilitates training given the limited data in this domain, (2) enables exploration of different deep learning structures for each module, (3) leads to developing less complex modules that are simpler to analyze, and (4) provides flexibility, permitting use of single modules within the framework or use of other modeling or machine learning applications, such as probabilistic graphical models or support vector machines. Our modular approach helps us meet the requirements of minimum accuracy levels that are demanded by the context of different decision points within a multi-class classification scheme. Convolutional Neural Networks are trained for each module for each sub-task with more than 90% classification accuracies on validation data set, and achieved classification accuracy of 96% for the task of GBM vs LGG classification, 71% for further identifying the grade of LGG into Grade II or Grade III on independent data set coming from new patients from the multi-institutional repository. PMID:26958289

Transfer learning improves supervised image segmentation across imaging protocols.

PubMed

van Opbroek, Annegreet; Ikram, M Arfan; Vernooij, Meike W; de Bruijne, Marleen

2015-05-01

The variation between images obtained with different scanners or different imaging protocols presents a major challenge in automatic segmentation of biomedical images. This variation especially hampers the application of otherwise successful supervised-learning techniques which, in order to perform well, often require a large amount of labeled training data that is exactly representative of the target data. We therefore propose to use transfer learning for image segmentation. Transfer-learning techniques can cope with differences in distributions between training and target data, and therefore may improve performance over supervised learning for segmentation across scanners and scan protocols. We present four transfer classifiers that can train a classification scheme with only a small amount of representative training data, in addition to a larger amount of other training data with slightly different characteristics. The performance of the four transfer classifiers was compared to that of standard supervised classification on two magnetic resonance imaging brain-segmentation tasks with multi-site data: white matter, gray matter, and cerebrospinal fluid segmentation; and white-matter-/MS-lesion segmentation. The experiments showed that when there is only a small amount of representative training data available, transfer learning can greatly outperform common supervised-learning approaches, minimizing classification errors by up to 60%.

Decomposition and extraction: a new framework for visual classification.

PubMed

Fang, Yuqiang; Chen, Qiang; Sun, Lin; Dai, Bin; Yan, Shuicheng

2014-08-01

In this paper, we present a novel framework for visual classification based on hierarchical image decomposition and hybrid midlevel feature extraction. Unlike most midlevel feature learning methods, which focus on the process of coding or pooling, we emphasize that the mechanism of image composition also strongly influences the feature extraction. To effectively explore the image content for the feature extraction, we model a multiplicity feature representation mechanism through meaningful hierarchical image decomposition followed by a fusion step. In particularly, we first propose a new hierarchical image decomposition approach in which each image is decomposed into a series of hierarchical semantical components, i.e, the structure and texture images. Then, different feature extraction schemes can be adopted to match the decomposed structure and texture processes in a dissociative manner. Here, two schemes are explored to produce property related feature representations. One is based on a single-stage network over hand-crafted features and the other is based on a multistage network, which can learn features from raw pixels automatically. Finally, those multiple midlevel features are incorporated by solving a multiple kernel learning task. Extensive experiments are conducted on several challenging data sets for visual classification, and experimental results demonstrate the effectiveness of the proposed method.

Multiple Sparse Representations Classification

PubMed Central

Plenge, Esben; Klein, Stefan S.; Niessen, Wiro J.; Meijering, Erik

2015-01-01

Sparse representations classification (SRC) is a powerful technique for pixelwise classification of images and it is increasingly being used for a wide variety of image analysis tasks. The method uses sparse representation and learned redundant dictionaries to classify image pixels. In this empirical study we propose to further leverage the redundancy of the learned dictionaries to achieve a more accurate classifier. In conventional SRC, each image pixel is associated with a small patch surrounding it. Using these patches, a dictionary is trained for each class in a supervised fashion. Commonly, redundant/overcomplete dictionaries are trained and image patches are sparsely represented by a linear combination of only a few of the dictionary elements. Given a set of trained dictionaries, a new patch is sparse coded using each of them, and subsequently assigned to the class whose dictionary yields the minimum residual energy. We propose a generalization of this scheme. The method, which we call multiple sparse representations classification (mSRC), is based on the observation that an overcomplete, class specific dictionary is capable of generating multiple accurate and independent estimates of a patch belonging to the class. So instead of finding a single sparse representation of a patch for each dictionary, we find multiple, and the corresponding residual energies provides an enhanced statistic which is used to improve classification. We demonstrate the efficacy of mSRC for three example applications: pixelwise classification of texture images, lumen segmentation in carotid artery magnetic resonance imaging (MRI), and bifurcation point detection in carotid artery MRI. We compare our method with conventional SRC, K-nearest neighbor, and support vector machine classifiers. The results show that mSRC outperforms SRC and the other reference methods. In addition, we present an extensive evaluation of the effect of the main mSRC parameters: patch size, dictionary size, and sparsity level. PMID:26177106

Brain-Computer Interface Based on Generation of Visual Images

PubMed Central

Bobrov, Pavel; Frolov, Alexander; Cantor, Charles; Fedulova, Irina; Bakhnyan, Mikhail; Zhavoronkov, Alexander

2011-01-01

This paper examines the task of recognizing EEG patterns that correspond to performing three mental tasks: relaxation and imagining of two types of pictures: faces and houses. The experiments were performed using two EEG headsets: BrainProducts ActiCap and Emotiv EPOC. The Emotiv headset becomes widely used in consumer BCI application allowing for conducting large-scale EEG experiments in the future. Since classification accuracy significantly exceeded the level of random classification during the first three days of the experiment with EPOC headset, a control experiment was performed on the fourth day using ActiCap. The control experiment has shown that utilization of high-quality research equipment can enhance classification accuracy (up to 68% in some subjects) and that the accuracy is independent of the presence of EEG artifacts related to blinking and eye movement. This study also shows that computationally-inexpensive Bayesian classifier based on covariance matrix analysis yields similar classification accuracy in this problem as a more sophisticated Multi-class Common Spatial Patterns (MCSP) classifier. PMID:21695206

Convolutional Neural Network-Based Robot Navigation Using Uncalibrated Spherical Images †

PubMed Central

Ran, Lingyan; Zhang, Yanning; Zhang, Qilin; Yang, Tao

2017-01-01

Vision-based mobile robot navigation is a vibrant area of research with numerous algorithms having been developed, the vast majority of which either belong to the scene-oriented simultaneous localization and mapping (SLAM) or fall into the category of robot-oriented lane-detection/trajectory tracking. These methods suffer from high computational cost and require stringent labelling and calibration efforts. To address these challenges, this paper proposes a lightweight robot navigation framework based purely on uncalibrated spherical images. To simplify the orientation estimation, path prediction and improve computational efficiency, the navigation problem is decomposed into a series of classification tasks. To mitigate the adverse effects of insufficient negative samples in the “navigation via classification” task, we introduce the spherical camera for scene capturing, which enables 360° fisheye panorama as training samples and generation of sufficient positive and negative heading directions. The classification is implemented as an end-to-end Convolutional Neural Network (CNN), trained on our proposed Spherical-Navi image dataset, whose category labels can be efficiently collected. This CNN is capable of predicting potential path directions with high confidence levels based on a single, uncalibrated spherical image. Experimental results demonstrate that the proposed framework outperforms competing ones in realistic applications. PMID:28604624

Classification

NASA Technical Reports Server (NTRS)

Oza, Nikunj C.

2011-01-01

A supervised learning task involves constructing a mapping from input data (normally described by several features) to the appropriate outputs. Within supervised learning, one type of task is a classification learning task, in which each output is one or more classes to which the input belongs. In supervised learning, a set of training examples---examples with known output values---is used by a learning algorithm to generate a model. This model is intended to approximate the mapping between the inputs and outputs. This model can be used to generate predicted outputs for inputs that have not been seen before. For example, we may have data consisting of observations of sunspots. In a classification learning task, our goal may be to learn to classify sunspots into one of several types. Each example may correspond to one candidate sunspot with various measurements or just an image. A learning algorithm would use the supplied examples to generate a model that approximates the mapping between each supplied set of measurements and the type of sunspot. This model can then be used to classify previously unseen sunspots based on the candidate's measurements. This chapter discusses methods to perform machine learning, with examples involving astronomy.

Crowdsourcing as a novel technique for retinal fundus photography classification: analysis of images in the EPIC Norfolk cohort on behalf of the UK Biobank Eye and Vision Consortium.

PubMed

Mitry, Danny; Peto, Tunde; Hayat, Shabina; Morgan, James E; Khaw, Kay-Tee; Foster, Paul J

2013-01-01

Crowdsourcing is the process of outsourcing numerous tasks to many untrained individuals. Our aim was to assess the performance and repeatability of crowdsourcing for the classification of retinal fundus photography. One hundred retinal fundus photograph images with pre-determined disease criteria were selected by experts from a large cohort study. After reading brief instructions and an example classification, we requested that knowledge workers (KWs) from a crowdsourcing platform classified each image as normal or abnormal with grades of severity. Each image was classified 20 times by different KWs. Four study designs were examined to assess the effect of varying incentive and KW experience in classification accuracy. All study designs were conducted twice to examine repeatability. Performance was assessed by comparing the sensitivity, specificity and area under the receiver operating characteristic curve (AUC). Without restriction on eligible participants, two thousand classifications of 100 images were received in under 24 hours at minimal cost. In trial 1 all study designs had an AUC (95%CI) of 0.701(0.680-0.721) or greater for classification of normal/abnormal. In trial 1, the highest AUC (95%CI) for normal/abnormal classification was 0.757 (0.738-0.776) for KWs with moderate experience. Comparable results were observed in trial 2. In trial 1, between 64-86% of any abnormal image was correctly classified by over half of all KWs. In trial 2, this ranged between 74-97%. Sensitivity was ≥ 96% for normal versus severely abnormal detections across all trials. Sensitivity for normal versus mildly abnormal varied between 61-79% across trials. With minimal training, crowdsourcing represents an accurate, rapid and cost-effective method of retinal image analysis which demonstrates good repeatability. Larger studies with more comprehensive participant training are needed to explore the utility of this compelling technique in large scale medical image analysis.

Multi-Modality Cascaded Convolutional Neural Networks for Alzheimer's Disease Diagnosis.

PubMed

Liu, Manhua; Cheng, Danni; Wang, Kundong; Wang, Yaping

2018-03-23

Accurate and early diagnosis of Alzheimer's disease (AD) plays important role for patient care and development of future treatment. Structural and functional neuroimages, such as magnetic resonance images (MRI) and positron emission tomography (PET), are providing powerful imaging modalities to help understand the anatomical and functional neural changes related to AD. In recent years, machine learning methods have been widely studied on analysis of multi-modality neuroimages for quantitative evaluation and computer-aided-diagnosis (CAD) of AD. Most existing methods extract the hand-craft imaging features after image preprocessing such as registration and segmentation, and then train a classifier to distinguish AD subjects from other groups. This paper proposes to construct cascaded convolutional neural networks (CNNs) to learn the multi-level and multimodal features of MRI and PET brain images for AD classification. First, multiple deep 3D-CNNs are constructed on different local image patches to transform the local brain image into more compact high-level features. Then, an upper high-level 2D-CNN followed by softmax layer is cascaded to ensemble the high-level features learned from the multi-modality and generate the latent multimodal correlation features of the corresponding image patches for classification task. Finally, these learned features are combined by a fully connected layer followed by softmax layer for AD classification. The proposed method can automatically learn the generic multi-level and multimodal features from multiple imaging modalities for classification, which are robust to the scale and rotation variations to some extent. No image segmentation and rigid registration are required in pre-processing the brain images. Our method is evaluated on the baseline MRI and PET images of 397 subjects including 93 AD patients, 204 mild cognitive impairment (MCI, 76 pMCI +128 sMCI) and 100 normal controls (NC) from Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Experimental results show that the proposed method achieves an accuracy of 93.26% for classification of AD vs. NC and 82.95% for classification pMCI vs. NC, demonstrating the promising classification performance.

A comparative study of deep learning models for medical image classification

NASA Astrophysics Data System (ADS)

Dutta, Suvajit; Manideep, B. C. S.; Rai, Shalva; Vijayarajan, V.

2017-11-01

Deep Learning(DL) techniques are conquering over the prevailing traditional approaches of neural network, when it comes to the huge amount of dataset, applications requiring complex functions demanding increase accuracy with lower time complexities. Neurosciences has already exploited DL techniques, thus portrayed itself as an inspirational source for researchers exploring the domain of Machine learning. DL enthusiasts cover the areas of vision, speech recognition, motion planning and NLP as well, moving back and forth among fields. This concerns with building models that can successfully solve variety of tasks requiring intelligence and distributed representation. The accessibility to faster CPUs, introduction of GPUs-performing complex vector and matrix computations, supported agile connectivity to network. Enhanced software infrastructures for distributed computing worked in strengthening the thought that made researchers suffice DL methodologies. The paper emphases on the following DL procedures to traditional approaches which are performed manually for classifying medical images. The medical images are used for the study Diabetic Retinopathy(DR) and computed tomography (CT) emphysema data. Both DR and CT data diagnosis is difficult task for normal image classification methods. The initial work was carried out with basic image processing along with K-means clustering for identification of image severity levels. After determining image severity levels ANN has been applied on the data to get the basic classification result, then it is compared with the result of DNNs (Deep Neural Networks), which performed efficiently because of its multiple hidden layer features basically which increases accuracy factors, but the problem of vanishing gradient in DNNs made to consider Convolution Neural Networks (CNNs) as well for better results. The CNNs are found to be providing better outcomes when compared to other learning models aimed at classification of images. CNNs are favoured as they provide better visual processing models successfully classifying the noisy data as well. The work centres on the detection on Diabetic Retinopathy-loss in vision and recognition of computed tomography (CT) emphysema data measuring the severity levels for both cases. The paper discovers how various Machine Learning algorithms can be implemented ensuing a supervised approach, so as to get accurate results with less complexity possible.

CIFAR10-DVS: An Event-Stream Dataset for Object Classification

PubMed Central

Li, Hongmin; Liu, Hanchao; Ji, Xiangyang; Li, Guoqi; Shi, Luping

2017-01-01

Neuromorphic vision research requires high-quality and appropriately challenging event-stream datasets to support continuous improvement of algorithms and methods. However, creating event-stream datasets is a time-consuming task, which needs to be recorded using the neuromorphic cameras. Currently, there are limited event-stream datasets available. In this work, by utilizing the popular computer vision dataset CIFAR-10, we converted 10,000 frame-based images into 10,000 event streams using a dynamic vision sensor (DVS), providing an event-stream dataset of intermediate difficulty in 10 different classes, named as “CIFAR10-DVS.” The conversion of event-stream dataset was implemented by a repeated closed-loop smooth (RCLS) movement of frame-based images. Unlike the conversion of frame-based images by moving the camera, the image movement is more realistic in respect of its practical applications. The repeated closed-loop image movement generates rich local intensity changes in continuous time which are quantized by each pixel of the DVS camera to generate events. Furthermore, a performance benchmark in event-driven object classification is provided based on state-of-the-art classification algorithms. This work provides a large event-stream dataset and an initial benchmark for comparison, which may boost algorithm developments in even-driven pattern recognition and object classification. PMID:28611582

CIFAR10-DVS: An Event-Stream Dataset for Object Classification.

PubMed

Li, Hongmin; Liu, Hanchao; Ji, Xiangyang; Li, Guoqi; Shi, Luping

2017-01-01

Neuromorphic vision research requires high-quality and appropriately challenging event-stream datasets to support continuous improvement of algorithms and methods. However, creating event-stream datasets is a time-consuming task, which needs to be recorded using the neuromorphic cameras. Currently, there are limited event-stream datasets available. In this work, by utilizing the popular computer vision dataset CIFAR-10, we converted 10,000 frame-based images into 10,000 event streams using a dynamic vision sensor (DVS), providing an event-stream dataset of intermediate difficulty in 10 different classes, named as "CIFAR10-DVS." The conversion of event-stream dataset was implemented by a repeated closed-loop smooth (RCLS) movement of frame-based images. Unlike the conversion of frame-based images by moving the camera, the image movement is more realistic in respect of its practical applications. The repeated closed-loop image movement generates rich local intensity changes in continuous time which are quantized by each pixel of the DVS camera to generate events. Furthermore, a performance benchmark in event-driven object classification is provided based on state-of-the-art classification algorithms. This work provides a large event-stream dataset and an initial benchmark for comparison, which may boost algorithm developments in even-driven pattern recognition and object classification.

Skimming Digits: Neuromorphic Classification of Spike-Encoded Images

PubMed Central

Cohen, Gregory K.; Orchard, Garrick; Leng, Sio-Hoi; Tapson, Jonathan; Benosman, Ryad B.; van Schaik, André

2016-01-01

The growing demands placed upon the field of computer vision have renewed the focus on alternative visual scene representations and processing paradigms. Silicon retinea provide an alternative means of imaging the visual environment, and produce frame-free spatio-temporal data. This paper presents an investigation into event-based digit classification using N-MNIST, a neuromorphic dataset created with a silicon retina, and the Synaptic Kernel Inverse Method (SKIM), a learning method based on principles of dendritic computation. As this work represents the first large-scale and multi-class classification task performed using the SKIM network, it explores different training patterns and output determination methods necessary to extend the original SKIM method to support multi-class problems. Making use of SKIM networks applied to real-world datasets, implementing the largest hidden layer sizes and simultaneously training the largest number of output neurons, the classification system achieved a best-case accuracy of 92.87% for a network containing 10,000 hidden layer neurons. These results represent the highest accuracies achieved against the dataset to date and serve to validate the application of the SKIM method to event-based visual classification tasks. Additionally, the study found that using a square pulse as the supervisory training signal produced the highest accuracy for most output determination methods, but the results also demonstrate that an exponential pattern is better suited to hardware implementations as it makes use of the simplest output determination method based on the maximum value. PMID:27199646

Multi-temporal and multi-source remote sensing image classification by nonlinear relative normalization

NASA Astrophysics Data System (ADS)

Tuia, Devis; Marcos, Diego; Camps-Valls, Gustau

2016-10-01

Remote sensing image classification exploiting multiple sensors is a very challenging problem: data from different modalities are affected by spectral distortions and mis-alignments of all kinds, and this hampers re-using models built for one image to be used successfully in other scenes. In order to adapt and transfer models across image acquisitions, one must be able to cope with datasets that are not co-registered, acquired under different illumination and atmospheric conditions, by different sensors, and with scarce ground references. Traditionally, methods based on histogram matching have been used. However, they fail when densities have very different shapes or when there is no corresponding band to be matched between the images. An alternative builds upon manifold alignment. Manifold alignment performs a multidimensional relative normalization of the data prior to product generation that can cope with data of different dimensionality (e.g. different number of bands) and possibly unpaired examples. Aligning data distributions is an appealing strategy, since it allows to provide data spaces that are more similar to each other, regardless of the subsequent use of the transformed data. In this paper, we study a methodology that aligns data from different domains in a nonlinear way through kernelization. We introduce the Kernel Manifold Alignment (KEMA) method, which provides a flexible and discriminative projection map, exploits only a few labeled samples (or semantic ties) in each domain, and reduces to solving a generalized eigenvalue problem. We successfully test KEMA in multi-temporal and multi-source very high resolution classification tasks, as well as on the task of making a model invariant to shadowing for hyperspectral imaging.

A novel end-to-end classifier using domain transferred deep convolutional neural networks for biomedical images.

PubMed

Pang, Shuchao; Yu, Zhezhou; Orgun, Mehmet A

2017-03-01

Highly accurate classification of biomedical images is an essential task in the clinical diagnosis of numerous medical diseases identified from those images. Traditional image classification methods combined with hand-crafted image feature descriptors and various classifiers are not able to effectively improve the accuracy rate and meet the high requirements of classification of biomedical images. The same also holds true for artificial neural network models directly trained with limited biomedical images used as training data or directly used as a black box to extract the deep features based on another distant dataset. In this study, we propose a highly reliable and accurate end-to-end classifier for all kinds of biomedical images via deep learning and transfer learning. We first apply domain transferred deep convolutional neural network for building a deep model; and then develop an overall deep learning architecture based on the raw pixels of original biomedical images using supervised training. In our model, we do not need the manual design of the feature space, seek an effective feature vector classifier or segment specific detection object and image patches, which are the main technological difficulties in the adoption of traditional image classification methods. Moreover, we do not need to be concerned with whether there are large training sets of annotated biomedical images, affordable parallel computing resources featuring GPUs or long times to wait for training a perfect deep model, which are the main problems to train deep neural networks for biomedical image classification as observed in recent works. With the utilization of a simple data augmentation method and fast convergence speed, our algorithm can achieve the best accuracy rate and outstanding classification ability for biomedical images. We have evaluated our classifier on several well-known public biomedical datasets and compared it with several state-of-the-art approaches. We propose a robust automated end-to-end classifier for biomedical images based on a domain transferred deep convolutional neural network model that shows a highly reliable and accurate performance which has been confirmed on several public biomedical image datasets. Copyright © 2017 Elsevier Ireland Ltd. All rights reserved.

Wavelength-Adaptive Dehazing Using Histogram Merging-Based Classification for UAV Images

PubMed Central

Yoon, Inhye; Jeong, Seokhwa; Jeong, Jaeheon; Seo, Doochun; Paik, Joonki

2015-01-01

Since incoming light to an unmanned aerial vehicle (UAV) platform can be scattered by haze and dust in the atmosphere, the acquired image loses the original color and brightness of the subject. Enhancement of hazy images is an important task in improving the visibility of various UAV images. This paper presents a spatially-adaptive dehazing algorithm that merges color histograms with consideration of the wavelength-dependent atmospheric turbidity. Based on the wavelength-adaptive hazy image acquisition model, the proposed dehazing algorithm consists of three steps: (i) image segmentation based on geometric classes; (ii) generation of the context-adaptive transmission map; and (iii) intensity transformation for enhancing a hazy UAV image. The major contribution of the research is a novel hazy UAV image degradation model by considering the wavelength of light sources. In addition, the proposed transmission map provides a theoretical basis to differentiate visually important regions from others based on the turbidity and merged classification results. PMID:25808767

A mathematical model of neuro-fuzzy approximation in image classification

NASA Astrophysics Data System (ADS)

Gopalan, Sasi; Pinto, Linu; Sheela, C.; Arun Kumar M., N.

2016-06-01

Image digitization and explosion of World Wide Web has made traditional search for image, an inefficient method for retrieval of required grassland image data from large database. For a given input query image Content-Based Image Retrieval (CBIR) system retrieves the similar images from a large database. Advances in technology has increased the use of grassland image data in diverse areas such has agriculture, art galleries, education, industry etc. In all the above mentioned diverse areas it is necessary to retrieve grassland image data efficiently from a large database to perform an assigned task and to make a suitable decision. A CBIR system based on grassland image properties and it uses the aid of a feed-forward back propagation neural network for an effective image retrieval is proposed in this paper. Fuzzy Memberships plays an important role in the input space of the proposed system which leads to a combined neural fuzzy approximation in image classification. The CBIR system with mathematical model in the proposed work gives more clarity about fuzzy-neuro approximation and the convergence of the image features in a grassland image.

Evolutionary pruning of transfer learned deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis.

PubMed

Samala, Ravi K; Chan, Heang-Ping; Hadjiiski, Lubomir M; Helvie, Mark A; Richter, Caleb; Cha, Kenny

2018-05-01

Deep learning models are highly parameterized, resulting in difficulty in inference and transfer learning for image recognition tasks. In this work, we propose a layered pathway evolution method to compress a deep convolutional neural network (DCNN) for classification of masses in digital breast tomosynthesis (DBT). The objective is to prune the number of tunable parameters while preserving the classification accuracy. In the first stage transfer learning, 19 632 augmented regions-of-interest (ROIs) from 2454 mass lesions on mammograms were used to train a pre-trained DCNN on ImageNet. In the second stage transfer learning, the DCNN was used as a feature extractor followed by feature selection and random forest classification. The pathway evolution was performed using genetic algorithm in an iterative approach with tournament selection driven by count-preserving crossover and mutation. The second stage was trained with 9120 DBT ROIs from 228 mass lesions using leave-one-case-out cross-validation. The DCNN was reduced by 87% in the number of neurons, 34% in the number of parameters, and 95% in the number of multiply-and-add operations required in the convolutional layers. The test AUC on 89 mass lesions from 94 independent DBT cases before and after pruning were 0.88 and 0.90, respectively, and the difference was not statistically significant (p  >  0.05). The proposed DCNN compression approach can reduce the number of required operations by 95% while maintaining the classification performance. The approach can be extended to other deep neural networks and imaging tasks where transfer learning is appropriate.

Evolutionary pruning of transfer learned deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis

NASA Astrophysics Data System (ADS)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir M.; Helvie, Mark A.; Richter, Caleb; Cha, Kenny

2018-05-01

Deep learning models are highly parameterized, resulting in difficulty in inference and transfer learning for image recognition tasks. In this work, we propose a layered pathway evolution method to compress a deep convolutional neural network (DCNN) for classification of masses in digital breast tomosynthesis (DBT). The objective is to prune the number of tunable parameters while preserving the classification accuracy. In the first stage transfer learning, 19 632 augmented regions-of-interest (ROIs) from 2454 mass lesions on mammograms were used to train a pre-trained DCNN on ImageNet. In the second stage transfer learning, the DCNN was used as a feature extractor followed by feature selection and random forest classification. The pathway evolution was performed using genetic algorithm in an iterative approach with tournament selection driven by count-preserving crossover and mutation. The second stage was trained with 9120 DBT ROIs from 228 mass lesions using leave-one-case-out cross-validation. The DCNN was reduced by 87% in the number of neurons, 34% in the number of parameters, and 95% in the number of multiply-and-add operations required in the convolutional layers. The test AUC on 89 mass lesions from 94 independent DBT cases before and after pruning were 0.88 and 0.90, respectively, and the difference was not statistically significant (p  >  0.05). The proposed DCNN compression approach can reduce the number of required operations by 95% while maintaining the classification performance. The approach can be extended to other deep neural networks and imaging tasks where transfer learning is appropriate.

Efficient HIK SVM learning for image classification.

PubMed

Wu, Jianxin

2012-10-01

Histograms are used in almost every aspect of image processing and computer vision, from visual descriptors to image representations. Histogram intersection kernel (HIK) and support vector machine (SVM) classifiers are shown to be very effective in dealing with histograms. This paper presents contributions concerning HIK SVM for image classification. First, we propose intersection coordinate descent (ICD), a deterministic and scalable HIK SVM solver. ICD is much faster than, and has similar accuracies to, general purpose SVM solvers and other fast HIK SVM training methods. We also extend ICD to the efficient training of a broader family of kernels. Second, we show an important empirical observation that ICD is not sensitive to the C parameter in SVM, and we provide some theoretical analyses to explain this observation. ICD achieves high accuracies in many problems, using its default parameters. This is an attractive property for practitioners, because many image processing tasks are too large to choose SVM parameters using cross-validation.

Convolutional neural network-based classification system design with compressed wireless sensor network images.

PubMed

Ahn, Jungmo; Park, JaeYeon; Park, Donghwan; Paek, Jeongyeup; Ko, JeongGil

2018-01-01

With the introduction of various advanced deep learning algorithms, initiatives for image classification systems have transitioned over from traditional machine learning algorithms (e.g., SVM) to Convolutional Neural Networks (CNNs) using deep learning software tools. A prerequisite in applying CNN to real world applications is a system that collects meaningful and useful data. For such purposes, Wireless Image Sensor Networks (WISNs), that are capable of monitoring natural environment phenomena using tiny and low-power cameras on resource-limited embedded devices, can be considered as an effective means of data collection. However, with limited battery resources, sending high-resolution raw images to the backend server is a burdensome task that has direct impact on network lifetime. To address this problem, we propose an energy-efficient pre- and post- processing mechanism using image resizing and color quantization that can significantly reduce the amount of data transferred while maintaining the classification accuracy in the CNN at the backend server. We show that, if well designed, an image in its highly compressed form can be well-classified with a CNN model trained in advance using adequately compressed data. Our evaluation using a real image dataset shows that an embedded device can reduce the amount of transmitted data by ∼71% while maintaining a classification accuracy of ∼98%. Under the same conditions, this process naturally reduces energy consumption by ∼71% compared to a WISN that sends the original uncompressed 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.

Peculiarities of use of ECOC and AdaBoost based classifiers for thematic processing of hyperspectral data

NASA Astrophysics Data System (ADS)

Dementev, A. O.; Dmitriev, E. V.; Kozoderov, V. V.; Egorov, V. D.

2017-10-01

Hyperspectral imaging is up-to-date promising technology widely applied for the accurate thematic mapping. The presence of a large number of narrow survey channels allows us to use subtle differences in spectral characteristics of objects and to make a more detailed classification than in the case of using standard multispectral data. The difficulties encountered in the processing of hyperspectral images are usually associated with the redundancy of spectral information which leads to the problem of the curse of dimensionality. Methods currently used for recognizing objects on multispectral and hyperspectral images are usually based on standard base supervised classification algorithms of various complexity. Accuracy of these algorithms can be significantly different depending on considered classification tasks. In this paper we study the performance of ensemble classification methods for the problem of classification of the forest vegetation. Error correcting output codes and boosting are tested on artificial data and real hyperspectral images. It is demonstrates, that boosting gives more significant improvement when used with simple base classifiers. The accuracy in this case in comparable the error correcting output code (ECOC) classifier with Gaussian kernel SVM base algorithm. However the necessity of boosting ECOC with Gaussian kernel SVM is questionable. It is demonstrated, that selected ensemble classifiers allow us to recognize forest species with high enough accuracy which can be compared with ground-based forest inventory data.

Learning the ideal observer for SKE detection tasks by use of convolutional neural networks (Cum Laude Poster Award)

NASA Astrophysics Data System (ADS)

Zhou, Weimin; Anastasio, Mark A.

2018-03-01

It has been advocated that task-based measures of image quality (IQ) should be employed to evaluate and optimize imaging systems. Task-based measures of IQ quantify the performance of an observer on a medically relevant task. The Bayesian Ideal Observer (IO), which employs complete statistical information of the object and noise, achieves the upper limit of the performance for a binary signal classification task. However, computing the IO performance is generally analytically intractable and can be computationally burdensome when Markov-chain Monte Carlo (MCMC) techniques are employed. In this paper, supervised learning with convolutional neural networks (CNNs) is employed to approximate the IO test statistics for a signal-known-exactly and background-known-exactly (SKE/BKE) binary detection task. The receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) are compared to those produced by the analytically computed IO. The advantages of the proposed supervised learning approach for approximating the IO are demonstrated.

Transfer learning for visual categorization: a survey.

PubMed

Shao, Ling; Zhu, Fan; Li, Xuelong

2015-05-01

Regular machine learning and data mining techniques study the training data for future inferences under a major assumption that the future data are within the same feature space or have the same distribution as the training data. However, due to the limited availability of human labeled training data, training data that stay in the same feature space or have the same distribution as the future data cannot be guaranteed to be sufficient enough to avoid the over-fitting problem. In real-world applications, apart from data in the target domain, related data in a different domain can also be included to expand the availability of our prior knowledge about the target future data. Transfer learning addresses such cross-domain learning problems by extracting useful information from data in a related domain and transferring them for being used in target tasks. In recent years, with transfer learning being applied to visual categorization, some typical problems, e.g., view divergence in action recognition tasks and concept drifting in image classification tasks, can be efficiently solved. In this paper, we survey state-of-the-art transfer learning algorithms in visual categorization applications, such as object recognition, image classification, and human action recognition.

An Iterative Inference Procedure Applying Conditional Random Fields for Simultaneous Classification of Land Cover and Land Use

NASA Astrophysics Data System (ADS)

Albert, L.; Rottensteiner, F.; Heipke, C.

2015-08-01

Land cover and land use exhibit strong contextual dependencies. We propose a novel approach for the simultaneous classification of land cover and land use, where semantic and spatial context is considered. The image sites for land cover and land use classification form a hierarchy consisting of two layers: a land cover layer and a land use layer. We apply Conditional Random Fields (CRF) at both layers. The layers differ with respect to the image entities corresponding to the nodes, the employed features and the classes to be distinguished. In the land cover layer, the nodes represent super-pixels; in the land use layer, the nodes correspond to objects from a geospatial database. Both CRFs model spatial dependencies between neighbouring image sites. The complex semantic relations between land cover and land use are integrated in the classification process by using contextual features. We propose a new iterative inference procedure for the simultaneous classification of land cover and land use, in which the two classification tasks mutually influence each other. This helps to improve the classification accuracy for certain classes. The main idea of this approach is that semantic context helps to refine the class predictions, which, in turn, leads to more expressive context information. Thus, potentially wrong decisions can be reversed at later stages. The approach is designed for input data based on aerial images. Experiments are carried out on a test site to evaluate the performance of the proposed method. We show the effectiveness of the iterative inference procedure and demonstrate that a smaller size of the super-pixels has a positive influence on the classification result.

Histology image analysis for carcinoma detection and grading

PubMed Central

He, Lei; Long, L. Rodney; Antani, Sameer; Thoma, George R.

2012-01-01

This paper presents an overview of the image analysis techniques in the domain of histopathology, specifically, for the objective of automated carcinoma detection and classification. As in other biomedical imaging areas such as radiology, many computer assisted diagnosis (CAD) systems have been implemented to aid histopathologists and clinicians in cancer diagnosis and research, which have been attempted to significantly reduce the labor and subjectivity of traditional manual intervention with histology images. The task of automated histology image analysis is usually not simple due to the unique characteristics of histology imaging, including the variability in image preparation techniques, clinical interpretation protocols, and the complex structures and very large size of the images themselves. In this paper we discuss those characteristics, provide relevant background information about slide preparation and interpretation, and review the application of digital image processing techniques to the field of histology image analysis. In particular, emphasis is given to state-of-the-art image segmentation methods for feature extraction and disease classification. Four major carcinomas of cervix, prostate, breast, and lung are selected to illustrate the functions and capabilities of existing CAD systems. PMID:22436890

Psychophysical Reverse Correlation with Multiple Response Alternatives

ERIC Educational Resources Information Center

Dai, Huanping; Micheyl, Christophe

2010-01-01

Psychophysical reverse-correlation methods such as the "classification image" technique provide a unique tool to uncover the internal representations and decision strategies of individual participants in perceptual tasks. Over the past 30 years, these techniques have gained increasing popularity among both visual and auditory psychophysicists.…

Tasked-based quantification of measurement utility for ex vivo multi-spectral Mueller polarimetry of the uterine cervix

NASA Astrophysics Data System (ADS)

Kupinski, Meredith; Rehbinder, Jean; Haddad, Huda; Deby, Stanislas; Vizet, Jérémy; Teig, Benjamin; Nazac, André; Pierangelo, Angelo; Moreau, François; Novikova, Tatiana

2017-07-01

Significant contrast in visible wavelength Mueller matrix images for healthy and pre-cancerous regions of excised cervical tissue is shown. A novel classification algorithm is used to compute a test statistic from a small patient population.

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.

Filtering and left ventricle segmentation of the fetal heart in ultrasound images

NASA Astrophysics Data System (ADS)

Vargas-Quintero, Lorena; Escalante-Ramírez, Boris

2013-11-01

In this paper, we propose to use filtering methods and a segmentation algorithm for the analysis of fetal heart in ultrasound images. Since noise speckle makes difficult the analysis of ultrasound images, the filtering process becomes a useful task in these types of applications. The filtering techniques consider in this work assume that the speckle noise is a random variable with a Rayleigh distribution. We use two multiresolution methods: one based on wavelet decomposition and the another based on the Hermite transform. The filtering process is used as way to strengthen the performance of the segmentation tasks. For the wavelet-based approach, a Bayesian estimator at subband level for pixel classification is employed. The Hermite method computes a mask to find those pixels that are corrupted by speckle. On the other hand, we picked out a method based on a deformable model or "snake" to evaluate the influence of the filtering techniques in the segmentation task of left ventricle in fetal echocardiographic images.

Analysis and classification of commercial ham slice images using directional fractal dimension features.

PubMed

Mendoza, Fernando; Valous, Nektarios A; Allen, Paul; Kenny, Tony A; Ward, Paddy; Sun, Da-Wen

2009-02-01

This paper presents a novel and non-destructive approach to the appearance characterization and classification of commercial pork, turkey and chicken ham slices. Ham slice images were modelled using directional fractal (DF(0°;45°;90°;135°)) dimensions and a minimum distance classifier was adopted to perform the classification task. Also, the role of different colour spaces and the resolution level of the images on DF analysis were investigated. This approach was applied to 480 wafer thin ham slices from four types of hams (120 slices per type): i.e., pork (cooked and smoked), turkey (smoked) and chicken (roasted). DF features were extracted from digitalized intensity images in greyscale, and R, G, B, L(∗), a(∗), b(∗), H, S, and V colour components for three image resolution levels (100%, 50%, and 25%). Simulation results show that in spite of the complexity and high variability in colour and texture appearance, the modelling of ham slice images with DF dimensions allows the capture of differentiating textural features between the four commercial ham types. Independent DF features entail better discrimination than that using the average of four directions. However, DF dimensions reveal a high sensitivity to colour channel, orientation and image resolution for the fractal analysis. The classification accuracy using six DF dimension features (a(90°)(∗),a(135°)(∗),H(0°),H(45°),S(0°),H(90°)) was 93.9% for training data and 82.2% for testing data.

Toolkits and Libraries for Deep Learning.

PubMed

Erickson, Bradley J; Korfiatis, Panagiotis; Akkus, Zeynettin; Kline, Timothy; Philbrick, Kenneth

2017-08-01

Deep learning is an important new area of machine learning which encompasses a wide range of neural network architectures designed to complete various tasks. In the medical imaging domain, example tasks include organ segmentation, lesion detection, and tumor classification. The most popular network architecture for deep learning for images is the convolutional neural network (CNN). Whereas traditional machine learning requires determination and calculation of features from which the algorithm learns, deep learning approaches learn the important features as well as the proper weighting of those features to make predictions for new data. In this paper, we will describe some of the libraries and tools that are available to aid in the construction and efficient execution of deep learning as applied to medical images.

Automating cell detection and classification in human brain fluorescent microscopy images using dictionary learning and sparse coding.

PubMed

Alegro, Maryana; Theofilas, Panagiotis; Nguy, Austin; Castruita, Patricia A; Seeley, William; Heinsen, Helmut; Ushizima, Daniela M; Grinberg, Lea T

2017-04-15

Immunofluorescence (IF) plays a major role in quantifying protein expression in situ and understanding cell function. It is widely applied in assessing disease mechanisms and in drug discovery research. Automation of IF analysis can transform studies using experimental cell models. However, IF analysis of postmortem human tissue relies mostly on manual interaction, often subjected to low-throughput and prone to error, leading to low inter and intra-observer reproducibility. Human postmortem brain samples challenges neuroscientists because of the high level of autofluorescence caused by accumulation of lipofuscin pigment during aging, hindering systematic analyses. We propose a method for automating cell counting and classification in IF microscopy of human postmortem brains. Our algorithm speeds up the quantification task while improving reproducibility. Dictionary learning and sparse coding allow for constructing improved cell representations using IF images. These models are input for detection and segmentation methods. Classification occurs by means of color distances between cells and a learned set. Our method successfully detected and classified cells in 49 human brain images. We evaluated our results regarding true positive, false positive, false negative, precision, recall, false positive rate and F1 score metrics. We also measured user-experience and time saved compared to manual countings. We compared our results to four open-access IF-based cell-counting tools available in the literature. Our method showed improved accuracy for all data samples. The proposed method satisfactorily detects and classifies cells from human postmortem brain IF images, with potential to be generalized for applications in other counting tasks. Copyright © 2017 Elsevier B.V. All rights reserved.

Quality Evaluation of Land-Cover Classification Using Convolutional Neural Network

NASA Astrophysics Data System (ADS)

Dang, Y.; Zhang, J.; Zhao, Y.; Luo, F.; Ma, W.; Yu, F.

2018-04-01

Land-cover classification is one of the most important products of earth observation, which focuses mainly on profiling the physical characters of the land surface with temporal and distribution attributes and contains the information of both natural and man-made coverage elements, such as vegetation, soil, glaciers, rivers, lakes, marsh wetlands and various man-made structures. In recent years, the amount of high-resolution remote sensing data has increased sharply. Accordingly, the volume of land-cover classification products increases, as well as the need to evaluate such frequently updated products that is a big challenge. Conventionally, the automatic quality evaluation of land-cover classification is made through pixel-based classifying algorithms, which lead to a much trickier task and consequently hard to keep peace with the required updating frequency. In this paper, we propose a novel quality evaluation approach for evaluating the land-cover classification by a scene classification method Convolutional Neural Network (CNN) model. By learning from remote sensing data, those randomly generated kernels that serve as filter matrixes evolved to some operators that has similar functions to man-crafted operators, like Sobel operator or Canny operator, and there are other kernels learned by the CNN model that are much more complex and can't be understood as existing filters. The method using CNN approach as the core algorithm serves quality-evaluation tasks well since it calculates a bunch of outputs which directly represent the image's membership grade to certain classes. An automatic quality evaluation approach for the land-cover DLG-DOM coupling data (DLG for Digital Line Graphic, DOM for Digital Orthophoto Map) will be introduced in this paper. The CNN model as an robustness method for image evaluation, then brought out the idea of an automatic quality evaluation approach for land-cover classification. Based on this experiment, new ideas of quality evaluation of DLG-DOM coupling land-cover classification or other kinds of labelled remote sensing data can be further studied.

Applying local binary patterns in image clustering problems

NASA Astrophysics Data System (ADS)

Skorokhod, Nikolai N.; Elizarov, Alexey I.

2017-11-01

Due to the fact that the cloudiness plays a critical role in the Earth radiative balance, the study of the distribution of different types of clouds and their movements is relevant. The main sources of such information are artificial satellites that provide data in the form of images. The most commonly used method of solving tasks of processing and classification of images of clouds is based on the description of texture features. The use of a set of local binary patterns is proposed to describe the texture image.

Feature Extraction and Machine Learning for the Classification of Brazilian Savannah Pollen Grains

PubMed Central

Souza, Junior Silva; da Silva, Gercina Gonçalves

2016-01-01

The classification of pollen species and types is an important task in many areas like forensic palynology, archaeological palynology and melissopalynology. This paper presents the first annotated image dataset for the Brazilian Savannah pollen types that can be used to train and test computer vision based automatic pollen classifiers. A first baseline human and computer performance for this dataset has been established using 805 pollen images of 23 pollen types. In order to access the computer performance, a combination of three feature extractors and four machine learning techniques has been implemented, fine tuned and tested. The results of these tests are also presented in this paper. PMID:27276196

Recognition and classification of colon cells applying the ensemble of classifiers.

PubMed

Kruk, M; Osowski, S; Koktysz, R

2009-02-01

The paper presents the application of an ensemble of classifiers for the recognition of colon cells on the basis of the microscope colon image. The solved task include: segmentation of the individual cells from the image using the morphological operations, the preprocessing stages, leading to the extraction of features, selection of the most important features, and the classification stage applying the classifiers arranged in the form of ensemble. The paper presents and discusses the results concerning the recognition of four most important colon cell types: eosinophylic granulocyte, neutrophilic granulocyte, lymphocyte and plasmocyte. The proposed system is able to recognize the cells with the accuracy comparable to the human expert (around 5% of discrepancy of both results).

Boosting bonsai trees for handwritten/printed text discrimination

NASA Astrophysics Data System (ADS)

Ricquebourg, Yann; Raymond, Christian; Poirriez, Baptiste; Lemaitre, Aurélie; Coüasnon, Bertrand

2013-12-01

Boosting over decision-stumps proved its efficiency in Natural Language Processing essentially with symbolic features, and its good properties (fast, few and not critical parameters, not sensitive to over-fitting) could be of great interest in the numeric world of pixel images. In this article we investigated the use of boosting over small decision trees, in image classification processing, for the discrimination of handwritten/printed text. Then, we conducted experiments to compare it to usual SVM-based classification revealing convincing results with very close performance, but with faster predictions and behaving far less as a black-box. Those promising results tend to make use of this classifier in more complex recognition tasks like multiclass problems.

Neural correlates of the number–size interference task in children

PubMed Central

Kaufmann, Liane; Koppelstaetter, Florian; Siedentopf, Christian; Haala, Ilka; Haberlandt, Edda; Zimmerhackl, Lothar-Bernd; Felber, Stefan; Ischebeck, Anja

2010-01-01

In this functional magnetic resonance imaging study, 17 children were asked to make numerical and physical magnitude classifications while ignoring the other stimulus dimension (number–size interference task). Digit pairs were either incongruent (3 8) or neutral (3 8). Generally, numerical magnitude interferes with font size (congruity effect). Moreover, relative to numerically adjacent digits far ones yield quicker responses (distance effect). Behaviourally, robust distance and congruity effects were observed in both tasks. imaging baselline contrasts revealed activations in frontal, parietal, occipital and cerebellar areas bilaterally. Different from results usually reported for adultssmaller distances activated frontal, but not (intra-)parietal areas in children. Congruity effects became significant only in physical comparisons. Thus, even with comparable behavioural performance, cerebral activation patterns may differ substantially between children and adults. PMID:16603917

Cell dynamic morphology classification using deep convolutional neural networks.

PubMed

Li, Heng; Pang, Fengqian; Shi, Yonggang; Liu, Zhiwen

2018-05-15

Cell morphology is often used as a proxy measurement of cell status to understand cell physiology. Hence, interpretation of cell dynamic morphology is a meaningful task in biomedical research. Inspired by the recent success of deep learning, we here explore the application of convolutional neural networks (CNNs) to cell dynamic morphology classification. An innovative strategy for the implementation of CNNs is introduced in this study. Mouse lymphocytes were collected to observe the dynamic morphology, and two datasets were thus set up to investigate the performances of CNNs. Considering the installation of deep learning, the classification problem was simplified from video data to image data, and was then solved by CNNs in a self-taught manner with the generated image data. CNNs were separately performed in three installation scenarios and compared with existing methods. Experimental results demonstrated the potential of CNNs in cell dynamic morphology classification, and validated the effectiveness of the proposed strategy. CNNs were successfully applied to the classification problem, and outperformed the existing methods in the classification accuracy. For the installation of CNNs, transfer learning was proved to be a promising scheme. © 2018 International Society for Advancement of Cytometry. © 2018 International Society for Advancement of Cytometry.

A scene-analysis approach to remote sensing. [San Francisco, California

NASA Technical Reports Server (NTRS)

Tenenbaum, J. M. (Principal Investigator); Fischler, M. A.; Wolf, H. C.

1978-01-01

The author has identified the following significant results. Geometric correspondance between a sensed image and a symbolic map is established in an initial stage of processing by adjusting parameters of a sensed model so that the image features predicted from the map optimally match corresponding features extracted from the sensed image. Information in the map is then used to constrain where to look in an image, what to look for, and how to interpret what is seen. For simple monitoring tasks involving multispectral classification, these constraints significantly reduce computation, simplify interpretation, and improve the utility of the resulting information. Previously intractable tasks requiring spatial and textural analysis may become straightforward in the context established by the map knowledge. The use of map-guided image analysis in monitoring the volume of water in a reservoir, the number of boxcars in a railyard, and the number of ships in a harbor is demonstrated.

A spectral-structural bag-of-features scene classifier for very high spatial resolution remote sensing imagery

NASA Astrophysics Data System (ADS)

Zhao, Bei; Zhong, Yanfei; Zhang, Liangpei

2016-06-01

Land-use classification of very high spatial resolution remote sensing (VHSR) imagery is one of the most challenging tasks in the field of remote sensing image processing. However, the land-use classification is hard to be addressed by the land-cover classification techniques, due to the complexity of the land-use scenes. Scene classification is considered to be one of the expected ways to address the land-use classification issue. The commonly used scene classification methods of VHSR imagery are all derived from the computer vision community that mainly deal with terrestrial image recognition. Differing from terrestrial images, VHSR images are taken by looking down with airborne and spaceborne sensors, which leads to the distinct light conditions and spatial configuration of land cover in VHSR imagery. Considering the distinct characteristics, two questions should be answered: (1) Which type or combination of information is suitable for the VHSR imagery scene classification? (2) Which scene classification algorithm is best for VHSR imagery? In this paper, an efficient spectral-structural bag-of-features scene classifier (SSBFC) is proposed to combine the spectral and structural information of VHSR imagery. SSBFC utilizes the first- and second-order statistics (the mean and standard deviation values, MeanStd) as the statistical spectral descriptor for the spectral information of the VHSR imagery, and uses dense scale-invariant feature transform (SIFT) as the structural feature descriptor. From the experimental results, the spectral information works better than the structural information, while the combination of the spectral and structural information is better than any single type of information. Taking the characteristic of the spatial configuration into consideration, SSBFC uses the whole image scene as the scope of the pooling operator, instead of the scope generated by a spatial pyramid (SP) commonly used in terrestrial image classification. The experimental results show that the whole image as the scope of the pooling operator performs better than the scope generated by SP. In addition, SSBFC codes and pools the spectral and structural features separately to avoid mutual interruption between the spectral and structural features. The coding vectors of spectral and structural features are then concatenated into a final coding vector. Finally, SSBFC classifies the final coding vector by support vector machine (SVM) with a histogram intersection kernel (HIK). Compared with the latest scene classification methods, the experimental results with three VHSR datasets demonstrate that the proposed SSBFC performs better than the other classification methods for VHSR image scenes.

Land use mapping from CBERS-2 images with open source tools by applying different classification algorithms

NASA Astrophysics Data System (ADS)

Sanhouse-García, Antonio J.; Rangel-Peraza, Jesús Gabriel; Bustos-Terrones, Yaneth; García-Ferrer, Alfonso; Mesas-Carrascosa, Francisco J.

2016-02-01

Land cover classification is often based on different characteristics between their classes, but with great homogeneity within each one of them. This cover is obtained through field work or by mean of processing satellite images. Field work involves high costs; therefore, digital image processing techniques have become an important alternative to perform this task. However, in some developing countries and particularly in Casacoima municipality in Venezuela, there is a lack of geographic information systems due to the lack of updated information and high costs in software license acquisition. This research proposes a low cost methodology to develop thematic mapping of local land use and types of coverage in areas with scarce resources. Thematic mapping was developed from CBERS-2 images and spatial information available on the network using open source tools. The supervised classification method per pixel and per region was applied using different classification algorithms and comparing them among themselves. Classification method per pixel was based on Maxver algorithms (maximum likelihood) and Euclidean distance (minimum distance), while per region classification was based on the Bhattacharya algorithm. Satisfactory results were obtained from per region classification, where overall reliability of 83.93% and kappa index of 0.81% were observed. Maxver algorithm showed a reliability value of 73.36% and kappa index 0.69%, while Euclidean distance obtained values of 67.17% and 0.61% for reliability and kappa index, respectively. It was demonstrated that the proposed methodology was very useful in cartographic processing and updating, which in turn serve as a support to develop management plans and land management. Hence, open source tools showed to be an economically viable alternative not only for forestry organizations, but for the general public, allowing them to develop projects in economically depressed and/or environmentally threatened areas.

EVALUATION OF REGISTRATION, COMPRESSION AND CLASSIFICATION ALGORITHMS

NASA Technical Reports Server (NTRS)

Jayroe, R. R.

1994-01-01

Several types of algorithms are generally used to process digital imagery such as Landsat data. The most commonly used algorithms perform the task of registration, compression, and classification. Because there are different techniques available for performing registration, compression, and classification, imagery data users need a rationale for selecting a particular approach to meet their particular needs. This collection of registration, compression, and classification algorithms was developed so that different approaches could be evaluated and the best approach for a particular application determined. Routines are included for six registration algorithms, six compression algorithms, and two classification algorithms. The package also includes routines for evaluating the effects of processing on the image data. This collection of routines should be useful to anyone using or developing image processing software. Registration of image data involves the geometrical alteration of the imagery. Registration routines available in the evaluation package include image magnification, mapping functions, partitioning, map overlay, and data interpolation. The compression of image data involves reducing the volume of data needed for a given image. Compression routines available in the package include adaptive differential pulse code modulation, two-dimensional transforms, clustering, vector reduction, and picture segmentation. Classification of image data involves analyzing the uncompressed or compressed image data to produce inventories and maps of areas of similar spectral properties within a scene. The classification routines available include a sequential linear technique and a maximum likelihood technique. The choice of the appropriate evaluation criteria is quite important in evaluating the image processing functions. The user is therefore given a choice of evaluation criteria with which to investigate the available image processing functions. All of the available evaluation criteria basically compare the observed results with the expected results. For the image reconstruction processes of registration and compression, the expected results are usually the original data or some selected characteristics of the original data. For classification processes the expected result is the ground truth of the scene. Thus, the comparison process consists of determining what changes occur in processing, where the changes occur, how much change occurs, and the amplitude of the change. The package includes evaluation routines for performing such comparisons as average uncertainty, average information transfer, chi-square statistics, multidimensional histograms, and computation of contingency matrices. This collection of routines is written in FORTRAN IV for batch execution and has been implemented on an IBM 360 computer with a central memory requirement of approximately 662K of 8 bit bytes. This collection of image processing and evaluation routines was developed in 1979.

A task-difficulty artifact in subliminal priming.

PubMed

Pratte, Michael S; Rouder, Jeffrey N

2009-08-01

Subliminal priming is said to occur when a subliminal prime influences the classification of a subsequent target. Most subliminal-priming claims are based on separate target- and prime-classification tasks. Because primes are intended to be subliminal, the prime-classification task is difficult, and the target-classification task is easy. To assess whether this task-difficulty difference accounts for previous claims of subliminal priming, we manipulated the ease of the prime-classification task by intermixing long-duration (visible) primes with short-duration (near liminal) ones. In Experiment 1, this strategy of intermixing long-duration primes raised classification of the short-duration ones. In Experiments 2 and 3, prime duration was lowered in such a way that prime classification was at chance in intermixed presentations. Under these conditions, we failed to observe any priming effects; hence, previous demonstrations of subliminal priming may simply have reflected a task-difficulty artifact.

Functional Heterogeneity and Convergence in the Right Temporoparietal Junction

PubMed Central

Lee, Su Mei; McCarthy, Gregory

2016-01-01

The right temporoparietal junction (rTPJ) is engaged by tasks that manipulate biological motion processing, Theory of Mind attributions, and attention reorienting. The proximity of activations elicited by these tasks raises the question of whether these tasks share common cognitive component processes that are subserved by common neural substrates. Here, we used high-resolution whole-brain functional magnetic resonance imaging in a within-subjects design to determine whether these tasks activate common regions of the rTPJ. Each participant was presented with the 3 tasks in the same imaging session. In a whole-brain analysis, we found that only the right and left TPJs were activated by all 3 tasks. Multivoxel pattern analysis revealed that the regions of overlap could still discriminate the 3 tasks. Notably, we found significant cross-task classification in the right TPJ, which suggests a shared neural process between the 3 tasks. Taken together, these results support prior studies that have indicated functional heterogeneity within the rTPJ but also suggest a convergence of function within a region of overlap. These results also call for further investigation into the nature of the function subserved in this overlap region. PMID:25477367

Reverse correlating love: highly passionate women idealize their partner's facial appearance.

PubMed

Gunaydin, Gul; DeLong, Jordan E

2015-01-01

A defining feature of passionate love is idealization--evaluating romantic partners in an overly favorable light. Although passionate love can be expected to color how favorably individuals represent their partner in their mind, little is known about how passionate love is linked with visual representations of the partner. Using reverse correlation techniques for the first time to study partner representations, the present study investigated whether women who are passionately in love represent their partner's facial appearance more favorably than individuals who are less passionately in love. In a within-participants design, heterosexual women completed two forced-choice classification tasks, one for their romantic partner and one for a male acquaintance, and a measure of passionate love. In each classification task, participants saw two faces superimposed with noise and selected the face that most resembled their partner (or an acquaintance). Classification images for each of high passion and low passion groups were calculated by averaging across noise patterns selected as resembling the partner or the acquaintance and superimposing the averaged noise on an average male face. A separate group of women evaluated the classification images on attractiveness, trustworthiness, and competence. Results showed that women who feel high (vs. low) passionate love toward their partner tend to represent his face as more attractive and trustworthy, even when controlling for familiarity effects using the acquaintance representation. Using an innovative method to study partner representations, these findings extend our understanding of cognitive processes in romantic relationships.

Reverse Correlating Love: Highly Passionate Women Idealize Their Partner’s Facial Appearance

PubMed Central

Gunaydin, Gul; DeLong, Jordan E.

2015-01-01

A defining feature of passionate love is idealization—evaluating romantic partners in an overly favorable light. Although passionate love can be expected to color how favorably individuals represent their partner in their mind, little is known about how passionate love is linked with visual representations of the partner. Using reverse correlation techniques for the first time to study partner representations, the present study investigated whether women who are passionately in love represent their partner’s facial appearance more favorably than individuals who are less passionately in love. In a within-participants design, heterosexual women completed two forced-choice classification tasks, one for their romantic partner and one for a male acquaintance, and a measure of passionate love. In each classification task, participants saw two faces superimposed with noise and selected the face that most resembled their partner (or an acquaintance). Classification images for each of high passion and low passion groups were calculated by averaging across noise patterns selected as resembling the partner or the acquaintance and superimposing the averaged noise on an average male face. A separate group of women evaluated the classification images on attractiveness, trustworthiness, and competence. Results showed that women who feel high (vs. low) passionate love toward their partner tend to represent his face as more attractive and trustworthy, even when controlling for familiarity effects using the acquaintance representation. Using an innovative method to study partner representations, these findings extend our understanding of cognitive processes in romantic relationships. PMID:25806540

Optical character recognition of camera-captured images based on phase features

NASA Astrophysics Data System (ADS)

Diaz-Escobar, Julia; Kober, Vitaly

2015-09-01

Nowadays most of digital information is obtained using mobile devices specially smartphones. In particular, it brings the opportunity for optical character recognition in camera-captured images. For this reason many recognition applications have been recently developed such as recognition of license plates, business cards, receipts and street signal; document classification, augmented reality, language translator and so on. Camera-captured images are usually affected by geometric distortions, nonuniform illumination, shadow, noise, which make difficult the recognition task with existing systems. It is well known that the Fourier phase contains a lot of important information regardless of the Fourier magnitude. So, in this work we propose a phase-based recognition system exploiting phase-congruency features for illumination/scale invariance. The performance of the proposed system is tested in terms of miss classifications and false alarms with the help of computer simulation.

Quality grading of Atlantic salmon (Salmo salar) by computer vision.

PubMed

Misimi, E; Erikson, U; Skavhaug, A

2008-06-01

In this study, we present a promising method of computer vision-based quality grading of whole Atlantic salmon (Salmo salar). Using computer vision, it was possible to differentiate among different quality grades of Atlantic salmon based on the external geometrical information contained in the fish images. Initially, before the image acquisition, the fish were subjectively graded and labeled into grading classes by a qualified human inspector in the processing plant. Prior to classification, the salmon images were segmented into binary images, and then feature extraction was performed on the geometrical parameters of the fish from the grading classes. The classification algorithm was a threshold-based classifier, which was designed using linear discriminant analysis. The performance of the classifier was tested by using the leave-one-out cross-validation method, and the classification results showed a good agreement between the classification done by human inspectors and by the computer vision. The computer vision-based method classified correctly 90% of the salmon from the data set as compared with the classification by human inspector. Overall, it was shown that computer vision can be used as a powerful tool to grade Atlantic salmon into quality grades in a fast and nondestructive manner by a relatively simple classifier algorithm. The low cost of implementation of today's advanced computer vision solutions makes this method feasible for industrial purposes in fish plants as it can replace manual labor, on which grading tasks still rely.

Homogenization of Classification Functions Measurement (HOCFUN): A Method for Measuring the Salience of Emotional Arousal in Thinking.

PubMed

Tonti, Marco; Salvatore, Sergio

2015-01-01

The problem of the measurement of emotion is a widely debated one. In this article we propose an instrument, the Homogenization of Classification Functions Measure (HOCFUN), designed for assessing the influence of emotional arousal on a rating task consisting of the evaluation of a sequence of images. The instrument defines an indicator (κ) that measures the degree of homogenization of the ratings given over 2 rating scales (pleasant-unpleasant and relevant-irrelevant). Such a degree of homogenization is interpreted as the effect of emotional arousal on thinking and therefore lends itself to be used as a marker of emotional arousal. A preliminary study of validation was implemented. The association of the κ indicator with 3 additional indicators was analyzed. Consistent with the hypotheses, the κ indicator proved to be associated, even if weakly and nonlinearly, with a marker of the homogenization of classification functions derived from a separate rating task and with 2 indirect indicators of emotional activation: the speed of performance on the HOCFUN task and an indicator of mood intensity. Taken as a whole, such results provide initial evidence supporting the HOCFUN construct validity.

Retrieval and classification of food images.

PubMed

Farinella, Giovanni Maria; Allegra, Dario; Moltisanti, Marco; Stanco, Filippo; Battiato, Sebastiano

2016-10-01

Automatic food understanding from images is an interesting challenge with applications in different domains. In particular, food intake monitoring is becoming more and more important because of the key role that it plays in health and market economies. In this paper, we address the study of food image processing from the perspective of Computer Vision. As first contribution we present a survey of the studies in the context of food image processing from the early attempts to the current state-of-the-art methods. Since retrieval and classification engines able to work on food images are required to build automatic systems for diet monitoring (e.g., to be embedded in wearable cameras), we focus our attention on the aspect of the representation of the food images because it plays a fundamental role in the understanding engines. The food retrieval and classification is a challenging task since the food presents high variableness and an intrinsic deformability. To properly study the peculiarities of different image representations we propose the UNICT-FD1200 dataset. It was composed of 4754 food images of 1200 distinct dishes acquired during real meals. Each food plate is acquired multiple times and the overall dataset presents both geometric and photometric variabilities. The images of the dataset have been manually labeled considering 8 categories: Appetizer, Main Course, Second Course, Single Course, Side Dish, Dessert, Breakfast, Fruit. We have performed tests employing different representations of the state-of-the-art to assess the related performances on the UNICT-FD1200 dataset. Finally, we propose a new representation based on the perceptual concept of Anti-Textons which is able to encode spatial information between Textons outperforming other representations in the context of food retrieval and Classification. Copyright © 2016 Elsevier Ltd. All rights reserved.

WE-G-207-05: Relationship Between CT Image Quality, Segmentation Performance, and Quantitative Image Feature Analysis

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

Lee, J; Nishikawa, R; Reiser, I

Purpose: Segmentation quality can affect quantitative image feature analysis. The objective of this study is to examine the relationship between computed tomography (CT) image quality, segmentation performance, and quantitative image feature analysis. Methods: A total of 90 pathology proven breast lesions in 87 dedicated breast CT images were considered. An iterative image reconstruction (IIR) algorithm was used to obtain CT images with different quality. With different combinations of 4 variables in the algorithm, this study obtained a total of 28 different qualities of CT images. Two imaging tasks/objectives were considered: 1) segmentation and 2) classification of the lesion as benignmore » or malignant. Twenty-three image features were extracted after segmentation using a semi-automated algorithm and 5 of them were selected via a feature selection technique. Logistic regression was trained and tested using leave-one-out-cross-validation and its area under the ROC curve (AUC) was recorded. The standard deviation of a homogeneous portion and the gradient of a parenchymal portion of an example breast were used as an estimate of image noise and sharpness. The DICE coefficient was computed using a radiologist’s drawing on the lesion. Mean DICE and AUC were used as performance metrics for each of the 28 reconstructions. The relationship between segmentation and classification performance under different reconstructions were compared. Distributions (median, 95% confidence interval) of DICE and AUC for each reconstruction were also compared. Results: Moderate correlation (Pearson’s rho = 0.43, p-value = 0.02) between DICE and AUC values was found. However, the variation between DICE and AUC values for each reconstruction increased as the image sharpness increased. There was a combination of IIR parameters that resulted in the best segmentation with the worst classification performance. Conclusion: There are certain images that yield better segmentation or classification performance. The best segmentation Result does not necessarily lead to the best classification Result. This work has been supported in part by grants from the NIH R21-EB015053. R Nishikawa is receives royalties form Hologic, Inc.« less

Using Bayesian neural networks to classify forest scenes

NASA Astrophysics Data System (ADS)

Vehtari, Aki; Heikkonen, Jukka; Lampinen, Jouko; Juujarvi, Jouni

1998-10-01

We present results that compare the performance of Bayesian learning methods for neural networks on the task of classifying forest scenes into trees and background. Classification task is demanding due to the texture richness of the trees, occlusions of the forest scene objects and diverse lighting conditions under operation. This makes it difficult to determine which are optimal image features for the classification. A natural way to proceed is to extract many different types of potentially suitable features, and to evaluate their usefulness in later processing stages. One approach to cope with large number of features is to use Bayesian methods to control the model complexity. Bayesian learning uses a prior on model parameters, combines this with evidence from a training data, and the integrates over the resulting posterior to make predictions. With this method, we can use large networks and many features without fear of overfitting. For this classification task we compare two Bayesian learning methods for multi-layer perceptron (MLP) neural networks: (1) The evidence framework of MacKay uses a Gaussian approximation to the posterior weight distribution and maximizes with respect to hyperparameters. (2) In a Markov Chain Monte Carlo (MCMC) method due to Neal, the posterior distribution of the network parameters is numerically integrated using the MCMC method. As baseline classifiers for comparison we use (3) MLP early stop committee, (4) K-nearest-neighbor and (5) Classification And Regression Tree.

Object based image analysis for the classification of the growth stages of Avocado crop, in Michoacán State, Mexico

NASA Astrophysics Data System (ADS)

Gao, Yan; Marpu, Prashanth; Morales Manila, Luis M.

2014-11-01

This paper assesses the suitability of 8-band Worldview-2 (WV2) satellite data and object-based random forest algorithm for the classification of avocado growth stages in Mexico. We tested both pixel-based with minimum distance (MD) and maximum likelihood (MLC) and object-based with Random Forest (RF) algorithm for this task. Training samples and verification data were selected by visual interpreting the WV2 images for seven thematic classes: fully grown, middle stage, and early stage of avocado crops, bare land, two types of natural forests, and water body. To examine the contribution of the four new spectral bands of WV2 sensor, all the tested classifications were carried out with and without the four new spectral bands. Classification accuracy assessment results show that object-based classification with RF algorithm obtained higher overall higher accuracy (93.06%) than pixel-based MD (69.37%) and MLC (64.03%) method. For both pixel-based and object-based methods, the classifications with the four new spectral bands (overall accuracy obtained higher accuracy than those without: overall accuracy of object-based RF classification with vs without: 93.06% vs 83.59%, pixel-based MD: 69.37% vs 67.2%, pixel-based MLC: 64.03% vs 36.05%, suggesting that the four new spectral bands in WV2 sensor contributed to the increase of the classification accuracy.

Multi-task transfer learning deep convolutional neural network: application to computer-aided diagnosis of breast cancer on mammograms

NASA Astrophysics Data System (ADS)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir M.; Helvie, Mark A.; Cha, Kenny H.; Richter, Caleb D.

2017-12-01

Transfer learning in deep convolutional neural networks (DCNNs) is an important step in its application to medical imaging tasks. We propose a multi-task transfer learning DCNN with the aim of translating the ‘knowledge’ learned from non-medical images to medical diagnostic tasks through supervised training and increasing the generalization capabilities of DCNNs by simultaneously learning auxiliary tasks. We studied this approach in an important application: classification of malignant and benign breast masses. With Institutional Review Board (IRB) approval, digitized screen-film mammograms (SFMs) and digital mammograms (DMs) were collected from our patient files and additional SFMs were obtained from the Digital Database for Screening Mammography. The data set consisted of 2242 views with 2454 masses (1057 malignant, 1397 benign). In single-task transfer learning, the DCNN was trained and tested on SFMs. In multi-task transfer learning, SFMs and DMs were used to train the DCNN, which was then tested on SFMs. N-fold cross-validation with the training set was used for training and parameter optimization. On the independent test set, the multi-task transfer learning DCNN was found to have significantly (p  =  0.007) higher performance compared to the single-task transfer learning DCNN. This study demonstrates that multi-task transfer learning may be an effective approach for training DCNN in medical imaging applications when training samples from a single modality are limited.

Medial temporal lobe contributions to short-term memory for faces

PubMed Central

Race, Elizabeth; LaRocque, Karen F.; Keane, Margaret M.; Verfaellie, Mieke

2015-01-01

The role of the medial temporal lobes (MTL) in short-term memory (STM) remains a matter of debate. While imaging studies commonly show hippocampal activation during short-delay memory tasks, evidence from amnesic patients with MTL lesions is mixed. It has been argued that apparent STM impairments in amnesia may reflect long-term memory (LTM) contributions to performance. We challenge this conclusion by demonstrating that MTL amnesic patients show impaired delayed matching-to-sample (DMS) for faces in a task that meets both a traditional delay-based and a recently proposed distractor-based criterion for classification as a STM task. In Experiment 1, we demonstrate that our face DMS task meets the proposed distractor-based criterion for STM classification, in that extensive processing of delay-period distractor stimuli disrupts performance of healthy individuals. In Experiment 2, MTL amnesic patients with lesions extending into anterior subhippocampal cortex, but not patients with lesions limited to the hippocampus, show impaired performance on this task without distraction at delays as short as 8s, within temporal range of delay-based STM classification, in the context of intact perceptual matching performance. Experiment 3 provides support for the hypothesis that STM for faces relies on configural processing by showing that the extent to which healthy participants’ performance is disrupted by interference depends on the configural demands of the distractor task. Together, these findings are consistent with the notion that the amnesic impairment in STM for faces reflects a deficit in configural processing associated with subhippocampal cortices and provide novel evidence that the MTL supports cognition beyond the LTM domain. PMID:23937185

Protein subcellular location pattern classification in cellular images using latent discriminative models.

PubMed

Li, Jieyue; Xiong, Liang; Schneider, Jeff; Murphy, Robert F

2012-06-15

Knowledge of the subcellular location of a protein is crucial for understanding its functions. The subcellular pattern of a protein is typically represented as the set of cellular components in which it is located, and an important task is to determine this set from microscope images. In this article, we address this classification problem using confocal immunofluorescence images from the Human Protein Atlas (HPA) project. The HPA contains images of cells stained for many proteins; each is also stained for three reference components, but there are many other components that are invisible. Given one such cell, the task is to classify the pattern type of the stained protein. We first randomly select local image regions within the cells, and then extract various carefully designed features from these regions. This region-based approach enables us to explicitly study the relationship between proteins and different cell components, as well as the interactions between these components. To achieve these two goals, we propose two discriminative models that extend logistic regression with structured latent variables. The first model allows the same protein pattern class to be expressed differently according to the underlying components in different regions. The second model further captures the spatial dependencies between the components within the same cell so that we can better infer these components. To learn these models, we propose a fast approximate algorithm for inference, and then use gradient-based methods to maximize the data likelihood. In the experiments, we show that the proposed models help improve the classification accuracies on synthetic data and real cellular images. The best overall accuracy we report in this article for classifying 942 proteins into 13 classes of patterns is about 84.6%, which to our knowledge is the best so far. In addition, the dependencies learned are consistent with prior knowledge of cell organization. http://murphylab.web.cmu.edu/software/.

Automatic diabetic retinopathy classification

NASA Astrophysics Data System (ADS)

Bravo, María. A.; Arbeláez, Pablo A.

2017-11-01

Diabetic retinopathy (DR) is a disease in which the retina is damaged due to augmentation in the blood pressure of small vessels. DR is the major cause of blindness for diabetics. It has been shown that early diagnosis can play a major role in prevention of visual loss and blindness. This work proposes a computer based approach for the detection of DR in back-of-the-eye images based on the use of convolutional neural networks (CNNs). Our CNN uses deep architectures to classify Back-of-the-eye Retinal Photographs (BRP) in 5 stages of DR. Our method combines several preprocessing images of BRP to obtain an ACA score of 50.5%. Furthermore, we explore subproblems by training a larger CNN of our main classification task.

Automating Cell Detection and Classification in Human Brain Fluorescent Microscopy Images Using Dictionary Learning and Sparse Coding

PubMed Central

Alegro, Maryana; Theofilas, Panagiotis; Nguy, Austin; Castruita, Patricia A.; Seeley, William; Heinsen, Helmut; Ushizima, Daniela M.

2017-01-01

Background Immunofluorescence (IF) plays a major role in quantifying protein expression in situ and understanding cell function. It is widely applied in assessing disease mechanisms and in drug discovery research. Automation of IF analysis can transform studies using experimental cell models. However, IF analysis of postmortem human tissue relies mostly on manual interaction, often subjected to low-throughput and prone to error, leading to low inter and intra-observer reproducibility. Human postmortem brain samples challenges neuroscientists because of the high level of autofluorescence caused by accumulation of lipofuscin pigment during aging, hindering systematic analyses. We propose a method for automating cell counting and classification in IF microscopy of human postmortem brains. Our algorithm speeds up the quantification task while improving reproducibility. New method Dictionary learning and sparse coding allow for constructing improved cell representations using IF images. These models are input for detection and segmentation methods. Classification occurs by means of color distances between cells and a learned set. Results Our method successfully detected and classified cells in 49 human brain images. We evaluated our results regarding true positive, false positive, false negative, precision, recall, false positive rate and F1 score metrics. We also measured user-experience and time saved compared to manual countings. Comparison with existing methods We compared our results to four open-access IF-based cell-counting tools available in the literature. Our method showed improved accuracy for all data samples. Conclusion The proposed method satisfactorily detects and classifies cells from human postmortem brain IF images, with potential to be generalized for applications in other counting tasks. PMID:28267565

Hybrid Brain-Computer Interface Techniques for Improved Classification Accuracy and Increased Number of Commands: A Review.

PubMed

Hong, Keum-Shik; Khan, Muhammad Jawad

2017-01-01

In this article, non-invasive hybrid brain-computer interface (hBCI) technologies for improving classification accuracy and increasing the number of commands are reviewed. Hybridization combining more than two modalities is a new trend in brain imaging and prosthesis control. Electroencephalography (EEG), due to its easy use and fast temporal resolution, is most widely utilized in combination with other brain/non-brain signal acquisition modalities, for instance, functional near infrared spectroscopy (fNIRS), electromyography (EMG), electrooculography (EOG), and eye tracker. Three main purposes of hybridization are to increase the number of control commands, improve classification accuracy and reduce the signal detection time. Currently, such combinations of EEG + fNIRS and EEG + EOG are most commonly employed. Four principal components (i.e., hardware, paradigm, classifiers, and features) relevant to accuracy improvement are discussed. In the case of brain signals, motor imagination/movement tasks are combined with cognitive tasks to increase active brain-computer interface (BCI) accuracy. Active and reactive tasks sometimes are combined: motor imagination with steady-state evoked visual potentials (SSVEP) and motor imagination with P300. In the case of reactive tasks, SSVEP is most widely combined with P300 to increase the number of commands. Passive BCIs, however, are rare. After discussing the hardware and strategies involved in the development of hBCI, the second part examines the approaches used to increase the number of control commands and to enhance classification accuracy. The future prospects and the extension of hBCI in real-time applications for daily life scenarios are provided.

Hybrid Brain–Computer Interface Techniques for Improved Classification Accuracy and Increased Number of Commands: A Review

PubMed Central

Hong, Keum-Shik; Khan, Muhammad Jawad

2017-01-01

In this article, non-invasive hybrid brain–computer interface (hBCI) technologies for improving classification accuracy and increasing the number of commands are reviewed. Hybridization combining more than two modalities is a new trend in brain imaging and prosthesis control. Electroencephalography (EEG), due to its easy use and fast temporal resolution, is most widely utilized in combination with other brain/non-brain signal acquisition modalities, for instance, functional near infrared spectroscopy (fNIRS), electromyography (EMG), electrooculography (EOG), and eye tracker. Three main purposes of hybridization are to increase the number of control commands, improve classification accuracy and reduce the signal detection time. Currently, such combinations of EEG + fNIRS and EEG + EOG are most commonly employed. Four principal components (i.e., hardware, paradigm, classifiers, and features) relevant to accuracy improvement are discussed. In the case of brain signals, motor imagination/movement tasks are combined with cognitive tasks to increase active brain–computer interface (BCI) accuracy. Active and reactive tasks sometimes are combined: motor imagination with steady-state evoked visual potentials (SSVEP) and motor imagination with P300. In the case of reactive tasks, SSVEP is most widely combined with P300 to increase the number of commands. Passive BCIs, however, are rare. After discussing the hardware and strategies involved in the development of hBCI, the second part examines the approaches used to increase the number of control commands and to enhance classification accuracy. The future prospects and the extension of hBCI in real-time applications for daily life scenarios are provided. PMID:28790910

3D multi-view convolutional neural networks for lung nodule classification

PubMed Central

Kang, Guixia; Hou, Beibei; Zhang, Ningbo

2017-01-01

The 3D convolutional neural network (CNN) is able to make full use of the spatial 3D context information of lung nodules, and the multi-view strategy has been shown to be useful for improving the performance of 2D CNN in classifying lung nodules. In this paper, we explore the classification of lung nodules using the 3D multi-view convolutional neural networks (MV-CNN) with both chain architecture and directed acyclic graph architecture, including 3D Inception and 3D Inception-ResNet. All networks employ the multi-view-one-network strategy. We conduct a binary classification (benign and malignant) and a ternary classification (benign, primary malignant and metastatic malignant) on Computed Tomography (CT) images from Lung Image Database Consortium and Image Database Resource Initiative database (LIDC-IDRI). All results are obtained via 10-fold cross validation. As regards the MV-CNN with chain architecture, results show that the performance of 3D MV-CNN surpasses that of 2D MV-CNN by a significant margin. Finally, a 3D Inception network achieved an error rate of 4.59% for the binary classification and 7.70% for the ternary classification, both of which represent superior results for the corresponding task. We compare the multi-view-one-network strategy with the one-view-one-network strategy. The results reveal that the multi-view-one-network strategy can achieve a lower error rate than the one-view-one-network strategy. PMID:29145492

Texture classification using non-Euclidean Minkowski dilation

NASA Astrophysics Data System (ADS)

Florindo, Joao B.; Bruno, Odemir M.

2018-03-01

This study presents a new method to extract meaningful descriptors of gray-scale texture images using Minkowski morphological dilation based on the Lp metric. The proposed approach is motivated by the success previously achieved by Bouligand-Minkowski fractal descriptors on texture classification. In essence, such descriptors are directly derived from the morphological dilation of a three-dimensional representation of the gray-level pixels using the classical Euclidean metric. In this way, we generalize the dilation for different values of p in the Lp metric (Euclidean is a particular case when p = 2) and obtain the descriptors from the cumulated distribution of the distance transform computed over the texture image. The proposed method is compared to other state-of-the-art approaches (such as local binary patterns and textons for example) in the classification of two benchmark data sets (UIUC and Outex). The proposed descriptors outperformed all the other approaches in terms of rate of images correctly classified. The interesting results suggest the potential of these descriptors in this type of task, with a wide range of possible applications to real-world problems.

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.

Sentinel-2 Level 2A Prototype Processor: Architecture, Algorithms And First Results

NASA Astrophysics Data System (ADS)

Muller-Wilm, Uwe; Louis, Jerome; Richter, Rudolf; Gascon, Ferran; Niezette, Marc

2013-12-01

Sen2Core is a prototype processor for Sentinel-2 Level 2A product processing and formatting. The processor is developed for and with ESA and performs the tasks of Atmospheric Correction and Scene Classification of Level 1C input data. Level 2A outputs are: Bottom-Of- Atmosphere (BOA) corrected reflectance images, Aerosol Optical Thickness-, Water Vapour-, Scene Classification maps and Quality indicators, including cloud and snow probabilities. The Level 2A Product Formatting performed by the processor follows the specification of the Level 1C User Product.

Combining features from ERP components in single-trial EEG for discriminating four-category visual objects.

PubMed

Wang, Changming; Xiong, Shi; Hu, Xiaoping; Yao, Li; Zhang, Jiacai

2012-10-01

Categorization of images containing visual objects can be successfully recognized using single-trial electroencephalograph (EEG) measured when subjects view images. Previous studies have shown that task-related information contained in event-related potential (ERP) components could discriminate two or three categories of object images. In this study, we investigated whether four categories of objects (human faces, buildings, cats and cars) could be mutually discriminated using single-trial EEG data. Here, the EEG waveforms acquired while subjects were viewing four categories of object images were segmented into several ERP components (P1, N1, P2a and P2b), and then Fisher linear discriminant analysis (Fisher-LDA) was used to classify EEG features extracted from ERP components. Firstly, we compared the classification results using features from single ERP components, and identified that the N1 component achieved the highest classification accuracies. Secondly, we discriminated four categories of objects using combining features from multiple ERP components, and showed that combination of ERP components improved four-category classification accuracies by utilizing the complementarity of discriminative information in ERP components. These findings confirmed that four categories of object images could be discriminated with single-trial EEG and could direct us to select effective EEG features for classifying visual objects.

Artificial neural network detects human uncertainty

NASA Astrophysics Data System (ADS)

Hramov, Alexander E.; Frolov, Nikita S.; Maksimenko, Vladimir A.; Makarov, Vladimir V.; Koronovskii, Alexey A.; Garcia-Prieto, Juan; Antón-Toro, Luis Fernando; Maestú, Fernando; Pisarchik, Alexander N.

2018-03-01

Artificial neural networks (ANNs) are known to be a powerful tool for data analysis. They are used in social science, robotics, and neurophysiology for solving tasks of classification, forecasting, pattern recognition, etc. In neuroscience, ANNs allow the recognition of specific forms of brain activity from multichannel EEG or MEG data. This makes the ANN an efficient computational core for brain-machine systems. However, despite significant achievements of artificial intelligence in recognition and classification of well-reproducible patterns of neural activity, the use of ANNs for recognition and classification of patterns in neural networks still requires additional attention, especially in ambiguous situations. According to this, in this research, we demonstrate the efficiency of application of the ANN for classification of human MEG trials corresponding to the perception of bistable visual stimuli with different degrees of ambiguity. We show that along with classification of brain states associated with multistable image interpretations, in the case of significant ambiguity, the ANN can detect an uncertain state when the observer doubts about the image interpretation. With the obtained results, we describe the possible application of ANNs for detection of bistable brain activity associated with difficulties in the decision-making process.

Conjunctive patches subspace learning with side information for collaborative image retrieval.

PubMed

Zhang, Lining; Wang, Lipo; Lin, Weisi

2012-08-01

Content-Based Image Retrieval (CBIR) has attracted substantial attention during the past few years for its potential practical applications to image management. A variety of Relevance Feedback (RF) schemes have been designed to bridge the semantic gap between the low-level visual features and the high-level semantic concepts for an image retrieval task. Various Collaborative Image Retrieval (CIR) schemes aim to utilize the user historical feedback log data with similar and dissimilar pairwise constraints to improve the performance of a CBIR system. However, existing subspace learning approaches with explicit label information cannot be applied for a CIR task, although the subspace learning techniques play a key role in various computer vision tasks, e.g., face recognition and image classification. In this paper, we propose a novel subspace learning framework, i.e., Conjunctive Patches Subspace Learning (CPSL) with side information, for learning an effective semantic subspace by exploiting the user historical feedback log data for a CIR task. The CPSL can effectively integrate the discriminative information of labeled log images, the geometrical information of labeled log images and the weakly similar information of unlabeled images together to learn a reliable subspace. We formally formulate this problem into a constrained optimization problem and then present a new subspace learning technique to exploit the user historical feedback log data. Extensive experiments on both synthetic data sets and a real-world image database demonstrate the effectiveness of the proposed scheme in improving the performance of a CBIR system by exploiting the user historical feedback log data.

Practical quantification of necrosis in histological whole-slide images.

PubMed

Homeyer, André; Schenk, Andrea; Arlt, Janine; Dahmen, Uta; Dirsch, Olaf; Hahn, Horst K

2013-06-01

Since the histological quantification of necrosis is a common task in medical research and practice, we evaluate different image analysis methods for quantifying necrosis in whole-slide images. In a practical usage scenario, we assess the impact of different classification algorithms and feature sets on both accuracy and computation time. We show how a well-chosen combination of multiresolution features and an efficient postprocessing step enables the accurate quantification necrosis in gigapixel images in less than a minute. The results are general enough to be applied to other areas of histological image analysis as well. Copyright © 2013 Elsevier Ltd. All rights reserved.

Medical image classification using spatial adjacent histogram based on adaptive local binary patterns.

PubMed

Liu, Dong; Wang, Shengsheng; Huang, Dezhi; Deng, Gang; Zeng, Fantao; Chen, Huiling

2016-05-01

Medical image recognition is an important task in both computer vision and computational biology. In the field of medical image classification, representing an image based on local binary patterns (LBP) descriptor has become popular. However, most existing LBP-based methods encode the binary patterns in a fixed neighborhood radius and ignore the spatial relationships among local patterns. The ignoring of the spatial relationships in the LBP will cause a poor performance in the process of capturing discriminative features for complex samples, such as medical images obtained by microscope. To address this problem, in this paper we propose a novel method to improve local binary patterns by assigning an adaptive neighborhood radius for each pixel. Based on these adaptive local binary patterns, we further propose a spatial adjacent histogram strategy to encode the micro-structures for image representation. An extensive set of evaluations are performed on four medical datasets which show that the proposed method significantly improves standard LBP and compares favorably with several other prevailing approaches. Copyright © 2016 Elsevier Ltd. All rights reserved.

The design and performance characteristics of a cellular logic 3-D image classification processor

NASA Astrophysics Data System (ADS)

Ankeney, L. A.

1981-04-01

The introduction of high resolution scanning laser radar systems which are capable of collecting range and reflectivity images, is predicted to significantly influence the development of processors capable of performing autonomous target classification tasks. Actively sensed range images are shown to be superior to passively collected infrared images in both image stability and information content. An illustrated tutorial introduces cellular logic (neighborhood) transformations and two and three dimensional erosion and dilation operations which are used for noise filters and geometric shape measurement. A unique 'cookbook' approach to selecting a sequence of neighborhood transformations suitable for object measurement is developed and related to false alarm rate and algorithm effectiveness measures. The cookbook design approach is used to develop an algorithm to classify objects based upon their 3-D geometrical features. A Monte Carlo performance analysis is used to demonstrate the utility of the design approach by characterizing the ability of the algorithm to classify randomly positioned three dimensional objects in the presence of additive noise, scale variations, and other forms of image distortion.

Fuzzy ontologies for semantic interpretation of remotely sensed images

NASA Astrophysics Data System (ADS)

Djerriri, Khelifa; Malki, Mimoun

2015-10-01

Object-based image classification consists in the assignment of object that share similar attributes to object categories. To perform such a task the remote sensing expert uses its personal knowledge, which is rarely formalized. Ontologies have been proposed as solution to represent domain knowledge agreed by domain experts in a formal and machine readable language. Classical ontology languages are not appropriate to deal with imprecision or vagueness in knowledge. Fortunately, Description Logics for the semantic web has been enhanced by various approaches to handle such knowledge. This paper presents the extension of the traditional ontology-based interpretation with fuzzy ontology of main land-cover classes in Landsat8-OLI scenes (vegetation, built-up areas, water bodies, shadow, clouds, forests) objects. A good classification of image objects was obtained and the results highlight the potential of the method to be replicated over time and space in the perspective of transferability of the procedure.

Textural features for image classification

NASA Technical Reports Server (NTRS)

Haralick, R. M.; Dinstein, I.; Shanmugam, K.

1973-01-01

Description of some easily computable textural features based on gray-tone spatial dependances, and illustration of their application in category-identification tasks of three different kinds of image data - namely, photomicrographs of five kinds of sandstones, 1:20,000 panchromatic aerial photographs of eight land-use categories, and ERTS multispectral imagery containing several land-use categories. Two kinds of decision rules are used - one for which the decision regions are convex polyhedra (a piecewise-linear decision rule), and one for which the decision regions are rectangular parallelpipeds (a min-max decision rule). In each experiment the data set was divided into two parts, a training set and a test set. Test set identification accuracy is 89% for the photomicrographs, 82% for the aerial photographic imagery, and 83% for the satellite imagery. These results indicate that the easily computable textural features probably have a general applicability for a wide variety of image-classification applications.

a Single-Exposure Dual-Energy Computed Radiography Technique for Improved Nodule Detection and Classification in Chest Imaging

NASA Astrophysics Data System (ADS)

Zink, Frank Edward

The detection and classification of pulmonary nodules is of great interest in chest radiography. Nodules are often indicative of primary cancer, and their detection is particularly important in asymptomatic patients. The ability to classify nodules as calcified or non-calcified is important because calcification is a positive indicator that the nodule is benign. Dual-energy methods offer the potential to improve both the detection and classification of nodules by allowing the formation of material-selective images. Tissue-selective images can improve detection by virtue of the elimination of obscuring rib structure. Bone -selective images are essentially calcium images, allowing classification of the nodule. A dual-energy technique is introduced which uses a computed radiography system to acquire dual-energy chest radiographs in a single-exposure. All aspects of the dual-energy technique are described, with particular emphasis on scatter-correction, beam-hardening correction, and noise-reduction algorithms. The adaptive noise-reduction algorithm employed improves material-selective signal-to-noise ratio by up to a factor of seven with minimal sacrifice in selectivity. A clinical comparison study is described, undertaken to compare the dual-energy technique to conventional chest radiography for the tasks of nodule detection and classification. Observer performance data were collected using the Free Response Observer Characteristic (FROC) method and the bi-normal Alternative FROC (AFROC) performance model. Results of the comparison study, analyzed using two common multiple observer statistical models, showed that the dual-energy technique was superior to conventional chest radiography for detection of nodules at a statistically significant level (p < .05). Discussion of the comparison study emphasizes the unique combination of data collection and analysis techniques employed, as well as the limitations of comparison techniques in the larger context of technology assessment.

Computationally efficient target classification in multispectral image data with Deep Neural Networks

NASA Astrophysics Data System (ADS)

Cavigelli, Lukas; Bernath, Dominic; Magno, Michele; Benini, Luca

2016-10-01

Detecting and classifying targets in video streams from surveillance cameras is a cumbersome, error-prone and expensive task. Often, the incurred costs are prohibitive for real-time monitoring. This leads to data being stored locally or transmitted to a central storage site for post-incident examination. The required communication links and archiving of the video data are still expensive and this setup excludes preemptive actions to respond to imminent threats. An effective way to overcome these limitations is to build a smart camera that analyzes the data on-site, close to the sensor, and transmits alerts when relevant video sequences are detected. Deep neural networks (DNNs) have come to outperform humans in visual classifications tasks and are also performing exceptionally well on other computer vision tasks. The concept of DNNs and Convolutional Networks (ConvNets) can easily be extended to make use of higher-dimensional input data such as multispectral data. We explore this opportunity in terms of achievable accuracy and required computational effort. To analyze the precision of DNNs for scene labeling in an urban surveillance scenario we have created a dataset with 8 classes obtained in a field experiment. We combine an RGB camera with a 25-channel VIS-NIR snapshot sensor to assess the potential of multispectral image data for target classification. We evaluate several new DNNs, showing that the spectral information fused together with the RGB frames can be used to improve the accuracy of the system or to achieve similar accuracy with a 3x smaller computation effort. We achieve a very high per-pixel accuracy of 99.1%. Even for scarcely occurring, but particularly interesting classes, such as cars, 75% of the pixels are labeled correctly with errors occurring only around the border of the objects. This high accuracy was obtained with a training set of only 30 labeled images, paving the way for fast adaptation to various application scenarios.

Optical beam classification using deep learning: a comparison with rule- and feature-based classification

NASA Astrophysics Data System (ADS)

Alom, Md. Zahangir; Awwal, Abdul A. S.; Lowe-Webb, Roger; Taha, Tarek M.

2017-08-01

Deep-learning methods are gaining popularity because of their state-of-the-art performance in image classification tasks. In this paper, we explore classification of laser-beam images from the National Ignition Facility (NIF) using a novel deeplearning approach. NIF is the world's largest, most energetic laser. It has nearly 40,000 optics that precisely guide, reflect, amplify, and focus 192 laser beams onto a fusion target. NIF utilizes four petawatt lasers called the Advanced Radiographic Capability (ARC) to produce backlighting X-ray illumination to capture implosion dynamics of NIF experiments with picosecond temporal resolution. In the current operational configuration, four independent short-pulse ARC beams are created and combined in a split-beam configuration in each of two NIF apertures at the entry of the pre-amplifier. The subaperture beams then propagate through the NIF beampath up to the ARC compressor. Each ARC beamlet is separately compressed with a dedicated set of four gratings and recombined as sub-apertures for transport to the parabola vessel, where the beams are focused using parabolic mirrors and pointed to the target. Small angular errors in the compressor gratings can cause the sub-aperture beams to diverge from one another and prevent accurate alignment through the transport section between the compressor and parabolic mirrors. This is an off-normal condition that must be detected and corrected. The goal of the off-normal check is to determine whether the ARC beamlets are sufficiently overlapped into a merged single spot or diverged into two distinct spots. Thus, the objective of the current work is three-fold: developing a simple algorithm to perform off-normal classification, exploring the use of Convolutional Neural Network (CNN) for the same task, and understanding the inter-relationship of the two approaches. The CNN recognition results are compared with other machine-learning approaches, such as Deep Neural Network (DNN) and Support Vector Machine (SVM). The experimental results show around 96% classification accuracy using CNN; the CNN approach also provides comparable recognition results compared to the present feature-based off-normal detection. The feature-based solution was developed to capture the expertise of a human expert in classifying the images. The misclassified results are further studied to explain the differences and discover any discrepancies or inconsistencies in current classification.

Numerical observer for atherosclerotic plaque classification in spectral computed tomography

PubMed Central

Lorsakul, Auranuch; Fakhri, Georges El; Worstell, William; Ouyang, Jinsong; Rakvongthai, Yothin; Laine, Andrew F.; Li, Quanzheng

2016-01-01

Abstract. Spectral computed tomography (SCT) generates better image quality than conventional computed tomography (CT). It has overcome several limitations for imaging atherosclerotic plaque. However, the literature evaluating the performance of SCT based on objective image assessment is very limited for the task of discriminating plaques. We developed a numerical-observer method and used it to assess performance on discrimination vulnerable-plaque features and compared the performance among multienergy CT (MECT), dual-energy CT (DECT), and conventional CT methods. Our numerical observer was designed to incorporate all spectral information and comprised two-processing stages. First, each energy-window domain was preprocessed by a set of localized channelized Hotelling observers (CHO). In this step, the spectral image in each energy bin was decorrelated using localized prewhitening and matched filtering with a set of Laguerre–Gaussian channel functions. Second, the series of the intermediate scores computed from all the CHOs were integrated by a Hotelling observer with an additional prewhitening and matched filter. The overall signal-to-noise ratio (SNR) and the area under the receiver operating characteristic curve (AUC) were obtained, yielding an overall discrimination performance metric. The performance of our new observer was evaluated for the particular binary classification task of differentiating between alternative plaque characterizations in carotid arteries. A clinically realistic model of signal variability was also included in our simulation of the discrimination tasks. The inclusion of signal variation is a key to applying the proposed observer method to spectral CT data. Hence, the task-based approaches based on the signal-known-exactly/background-known-exactly (SKE/BKE) framework and the clinical-relevant signal-known-statistically/background-known-exactly (SKS/BKE) framework were applied for analytical computation of figures of merit (FOM). Simulated data of a carotid-atherosclerosis patient were used to validate our methods. We used an extended cardiac-torso anthropomorphic digital phantom and three simulated plaque types (i.e., calcified plaque, fatty-mixed plaque, and iodine-mixed blood). The images were reconstructed using a standard filtered backprojection (FBP) algorithm for all the acquisition methods and were applied to perform two different discrimination tasks of: (1) calcified plaque versus fatty-mixed plaque and (2) calcified plaque versus iodine-mixed blood. MECT outperformed DECT and conventional CT systems for all cases of the SKE/BKE and SKS/BKE tasks (all p<0.01). On average of signal variability, MECT yielded the SNR improvements over other acquisition methods in the range of 46.8% to 65.3% (all p<0.01) for FBP-Ramp images and 53.2% to 67.7% (all p<0.01) for FBP-Hanning images for both identification tasks. This proposed numerical observer combined with our signal variability framework is promising for assessing material characterization obtained through the additional energy-dependent attenuation information of SCT. These methods can be further extended to other clinical tasks such as kidney or urinary stone identification applications. PMID:27429999

Classifying magnetic resonance image modalities with convolutional neural networks

NASA Astrophysics Data System (ADS)

Remedios, Samuel; Pham, Dzung L.; Butman, John A.; Roy, Snehashis

2018-02-01

Magnetic Resonance (MR) imaging allows the acquisition of images with different contrast properties depending on the acquisition protocol and the magnetic properties of tissues. Many MR brain image processing techniques, such as tissue segmentation, require multiple MR contrasts as inputs, and each contrast is treated differently. Thus it is advantageous to automate the identification of image contrasts for various purposes, such as facilitating image processing pipelines, and managing and maintaining large databases via content-based image retrieval (CBIR). Most automated CBIR techniques focus on a two-step process: extracting features from data and classifying the image based on these features. We present a novel 3D deep convolutional neural network (CNN)- based method for MR image contrast classification. The proposed CNN automatically identifies the MR contrast of an input brain image volume. Specifically, we explored three classification problems: (1) identify T1-weighted (T1-w), T2-weighted (T2-w), and fluid-attenuated inversion recovery (FLAIR) contrasts, (2) identify pre vs postcontrast T1, (3) identify pre vs post-contrast FLAIR. A total of 3418 image volumes acquired from multiple sites and multiple scanners were used. To evaluate each task, the proposed model was trained on 2137 images and tested on the remaining 1281 images. Results showed that image volumes were correctly classified with 97.57% accuracy.

Digital mammographic tumor classification using transfer learning from deep convolutional neural networks.

PubMed

Huynh, Benjamin Q; Li, Hui; Giger, Maryellen L

2016-07-01

Convolutional neural networks (CNNs) show potential for computer-aided diagnosis (CADx) by learning features directly from the image data instead of using analytically extracted features. However, CNNs are difficult to train from scratch for medical images due to small sample sizes and variations in tumor presentations. Instead, transfer learning can be used to extract tumor information from medical images via CNNs originally pretrained for nonmedical tasks, alleviating the need for large datasets. Our database includes 219 breast lesions (607 full-field digital mammographic images). We compared support vector machine classifiers based on the CNN-extracted image features and our prior computer-extracted tumor features in the task of distinguishing between benign and malignant breast lesions. Five-fold cross validation (by lesion) was conducted with the area under the receiver operating characteristic (ROC) curve as the performance metric. Results show that classifiers based on CNN-extracted features (with transfer learning) perform comparably to those using analytically extracted features [area under the ROC curve [Formula: see text

Towards brain-activity-controlled information retrieval: Decoding image relevance from MEG signals.

PubMed

Kauppi, Jukka-Pekka; Kandemir, Melih; Saarinen, Veli-Matti; Hirvenkari, Lotta; Parkkonen, Lauri; Klami, Arto; Hari, Riitta; Kaski, Samuel

2015-05-15

We hypothesize that brain activity can be used to control future information retrieval systems. To this end, we conducted a feasibility study on predicting the relevance of visual objects from brain activity. We analyze both magnetoencephalographic (MEG) and gaze signals from nine subjects who were viewing image collages, a subset of which was relevant to a predetermined task. We report three findings: i) the relevance of an image a subject looks at can be decoded from MEG signals with performance significantly better than chance, ii) fusion of gaze-based and MEG-based classifiers significantly improves the prediction performance compared to using either signal alone, and iii) non-linear classification of the MEG signals using Gaussian process classifiers outperforms linear classification. These findings break new ground for building brain-activity-based interactive image retrieval systems, as well as for systems utilizing feedback both from brain activity and eye movements. Copyright © 2015 Elsevier Inc. All rights reserved.

Recent Advances of Malaria Parasites Detection Systems Based on Mathematical Morphology

PubMed Central

Di Ruberto, Cecilia; Kocher, Michel

2018-01-01

Malaria is an epidemic health disease and a rapid, accurate diagnosis is necessary for proper intervention. Generally, pathologists visually examine blood stained slides for malaria diagnosis. Nevertheless, this kind of visual inspection is subjective, error-prone and time-consuming. In order to overcome the issues, numerous methods of automatic malaria diagnosis have been proposed so far. In particular, many researchers have used mathematical morphology as a powerful tool for computer aided malaria detection and classification. Mathematical morphology is not only a theory for the analysis of spatial structures, but also a very powerful technique widely used for image processing purposes and employed successfully in biomedical image analysis, especially in preprocessing and segmentation tasks. Microscopic image analysis and particularly malaria detection and classification can greatly benefit from the use of morphological operators. The aim of this paper is to present a review of recent mathematical morphology based methods for malaria parasite detection and identification in stained blood smears images. PMID:29419781

Recent Advances of Malaria Parasites Detection Systems Based on Mathematical Morphology.

PubMed

Loddo, Andrea; Di Ruberto, Cecilia; Kocher, Michel

2018-02-08

Malaria is an epidemic health disease and a rapid, accurate diagnosis is necessary for proper intervention. Generally, pathologists visually examine blood stained slides for malaria diagnosis. Nevertheless, this kind of visual inspection is subjective, error-prone and time-consuming. In order to overcome the issues, numerous methods of automatic malaria diagnosis have been proposed so far. In particular, many researchers have used mathematical morphology as a powerful tool for computer aided malaria detection and classification. Mathematical morphology is not only a theory for the analysis of spatial structures, but also a very powerful technique widely used for image processing purposes and employed successfully in biomedical image analysis, especially in preprocessing and segmentation tasks. Microscopic image analysis and particularly malaria detection and classification can greatly benefit from the use of morphological operators. The aim of this paper is to present a review of recent mathematical morphology based methods for malaria parasite detection and identification in stained blood smears images.

Do you see what I see? Mobile eye-tracker contextual analysis and inter-rater reliability.

PubMed

Stuart, S; Hunt, D; Nell, J; Godfrey, A; Hausdorff, J M; Rochester, L; Alcock, L

2018-02-01

Mobile eye-trackers are currently used during real-world tasks (e.g. gait) to monitor visual and cognitive processes, particularly in ageing and Parkinson's disease (PD). However, contextual analysis involving fixation locations during such tasks is rarely performed due to its complexity. This study adapted a validated algorithm and developed a classification method to semi-automate contextual analysis of mobile eye-tracking data. We further assessed inter-rater reliability of the proposed classification method. A mobile eye-tracker recorded eye-movements during walking in five healthy older adult controls (HC) and five people with PD. Fixations were identified using a previously validated algorithm, which was adapted to provide still images of fixation locations (n = 116). The fixation location was manually identified by two raters (DH, JN), who classified the locations. Cohen's kappa correlation coefficients determined the inter-rater reliability. The algorithm successfully provided still images for each fixation, allowing manual contextual analysis to be performed. The inter-rater reliability for classifying the fixation location was high for both PD (kappa = 0.80, 95% agreement) and HC groups (kappa = 0.80, 91% agreement), which indicated a reliable classification method. This study developed a reliable semi-automated contextual analysis method for gait studies in HC and PD. Future studies could adapt this methodology for various gait-related eye-tracking studies.

Comparing the quality of accessing medical literature using content-based visual and textual information retrieval

NASA Astrophysics Data System (ADS)

Müller, Henning; Kalpathy-Cramer, Jayashree; Kahn, Charles E., Jr.; Hersh, William

2009-02-01

Content-based visual information (or image) retrieval (CBIR) has been an extremely active research domain within medical imaging over the past ten years, with the goal of improving the management of visual medical information. Many technical solutions have been proposed, and application scenarios for image retrieval as well as image classification have been set up. However, in contrast to medical information retrieval using textual methods, visual retrieval has only rarely been applied in clinical practice. This is despite the large amount and variety of visual information produced in hospitals every day. This information overload imposes a significant burden upon clinicians, and CBIR technologies have the potential to help the situation. However, in order for CBIR to become an accepted clinical tool, it must demonstrate a higher level of technical maturity than it has to date. Since 2004, the ImageCLEF benchmark has included a task for the comparison of visual information retrieval algorithms for medical applications. In 2005, a task for medical image classification was introduced and both tasks have been run successfully for the past four years. These benchmarks allow an annual comparison of visual retrieval techniques based on the same data sets and the same query tasks, enabling the meaningful comparison of various retrieval techniques. The datasets used from 2004-2007 contained images and annotations from medical teaching files. In 2008, however, the dataset used was made up of 67,000 images (along with their associated figure captions and the full text of their corresponding articles) from two Radiological Society of North America (RSNA) scientific journals. This article describes the results of the medical image retrieval task of the ImageCLEF 2008 evaluation campaign. We compare the retrieval results of both visual and textual information retrieval systems from 15 research groups on the aforementioned data set. The results show clearly that, currently, visual retrieval alone does not achieve the performance necessary for real-world clinical applications. Most of the common visual retrieval techniques have a MAP (Mean Average Precision) of around 2-3%, which is much lower than that achieved using textual retrieval (MAP=29%). Advanced machine learning techniques, together with good training data, have been shown to improve the performance of visual retrieval systems in the past. Multimodal retrieval (basing retrieval on both visual and textual information) can achieve better results than purely visual, but only when carefully applied. In many cases, multimodal retrieval systems performed even worse than purely textual retrieval systems. On the other hand, some multimodal retrieval systems demonstrated significantly increased early precision, which has been shown to be a desirable behavior in real-world systems.

Ensemble Classification of Alzheimer's Disease and Mild Cognitive Impairment Based on Complex Graph Measures from Diffusion Tensor Images

PubMed Central

Ebadi, Ashkan; Dalboni da Rocha, Josué L.; Nagaraju, Dushyanth B.; Tovar-Moll, Fernanda; Bramati, Ivanei; Coutinho, Gabriel; Sitaram, Ranganatha; Rashidi, Parisa

2017-01-01

The human brain is a complex network of interacting regions. The gray matter regions of brain are interconnected by white matter tracts, together forming one integrative complex network. In this article, we report our investigation about the potential of applying brain connectivity patterns as an aid in diagnosing Alzheimer's disease and Mild Cognitive Impairment (MCI). We performed pattern analysis of graph theoretical measures derived from Diffusion Tensor Imaging (DTI) data representing structural brain networks of 45 subjects, consisting of 15 patients of Alzheimer's disease (AD), 15 patients of MCI, and 15 healthy subjects (CT). We considered pair-wise class combinations of subjects, defining three separate classification tasks, i.e., AD-CT, AD-MCI, and CT-MCI, and used an ensemble classification module to perform the classification tasks. Our ensemble framework with feature selection shows a promising performance with classification accuracy of 83.3% for AD vs. MCI, 80% for AD vs. CT, and 70% for MCI vs. CT. Moreover, our findings suggest that AD can be related to graph measures abnormalities at Brodmann areas in the sensorimotor cortex and piriform cortex. In this way, node redundancy coefficient and load centrality in the primary motor cortex were recognized as good indicators of AD in contrast to MCI. In general, load centrality, betweenness centrality, and closeness centrality were found to be the most relevant network measures, as they were the top identified features at different nodes. The present study can be regarded as a “proof of concept” about a procedure for the classification of MRI markers between AD dementia, MCI, and normal old individuals, due to the small and not well-defined groups of AD and MCI patients. Future studies with larger samples of subjects and more sophisticated patient exclusion criteria are necessary toward the development of a more precise technique for clinical diagnosis. PMID:28293162

Accurate crop classification using hierarchical genetic fuzzy rule-based systems

NASA Astrophysics Data System (ADS)

Topaloglou, Charalampos A.; Mylonas, Stelios K.; Stavrakoudis, Dimitris G.; Mastorocostas, Paris A.; Theocharis, John B.

2014-10-01

This paper investigates the effectiveness of an advanced classification system for accurate crop classification using very high resolution (VHR) satellite imagery. Specifically, a recently proposed genetic fuzzy rule-based classification system (GFRBCS) is employed, namely, the Hierarchical Rule-based Linguistic Classifier (HiRLiC). HiRLiC's model comprises a small set of simple IF-THEN fuzzy rules, easily interpretable by humans. One of its most important attributes is that its learning algorithm requires minimum user interaction, since the most important learning parameters affecting the classification accuracy are determined by the learning algorithm automatically. HiRLiC is applied in a challenging crop classification task, using a SPOT5 satellite image over an intensively cultivated area in a lake-wetland ecosystem in northern Greece. A rich set of higher-order spectral and textural features is derived from the initial bands of the (pan-sharpened) image, resulting in an input space comprising 119 features. The experimental analysis proves that HiRLiC compares favorably to other interpretable classifiers of the literature, both in terms of structural complexity and classification accuracy. Its testing accuracy was very close to that obtained by complex state-of-the-art classification systems, such as the support vector machines (SVM) and random forest (RF) classifiers. Nevertheless, visual inspection of the derived classification maps shows that HiRLiC is characterized by higher generalization properties, providing more homogeneous classifications that the competitors. Moreover, the runtime requirements for producing the thematic map was orders of magnitude lower than the respective for the competitors.

Automated Melanoma Recognition in Dermoscopy Images via Very Deep Residual Networks.

PubMed

Yu, Lequan; Chen, Hao; Dou, Qi; Qin, Jing; Heng, Pheng-Ann

2017-04-01

Automated melanoma recognition in dermoscopy images is a very challenging task due to the low contrast of skin lesions, the huge intraclass variation of melanomas, the high degree of visual similarity between melanoma and non-melanoma lesions, and the existence of many artifacts in the image. In order to meet these challenges, we propose a novel method for melanoma recognition by leveraging very deep convolutional neural networks (CNNs). Compared with existing methods employing either low-level hand-crafted features or CNNs with shallower architectures, our substantially deeper networks (more than 50 layers) can acquire richer and more discriminative features for more accurate recognition. To take full advantage of very deep networks, we propose a set of schemes to ensure effective training and learning under limited training data. First, we apply the residual learning to cope with the degradation and overfitting problems when a network goes deeper. This technique can ensure that our networks benefit from the performance gains achieved by increasing network depth. Then, we construct a fully convolutional residual network (FCRN) for accurate skin lesion segmentation, and further enhance its capability by incorporating a multi-scale contextual information integration scheme. Finally, we seamlessly integrate the proposed FCRN (for segmentation) and other very deep residual networks (for classification) to form a two-stage framework. This framework enables the classification network to extract more representative and specific features based on segmented results instead of the whole dermoscopy images, further alleviating the insufficiency of training data. The proposed framework is extensively evaluated on ISBI 2016 Skin Lesion Analysis Towards Melanoma Detection Challenge dataset. Experimental results demonstrate the significant performance gains of the proposed framework, ranking the first in classification and the second in segmentation among 25 teams and 28 teams, respectively. This study corroborates that very deep CNNs with effective training mechanisms can be employed to solve complicated medical image analysis tasks, even with limited training data.

Geographical topic learning for social images with a deep neural network

NASA Astrophysics Data System (ADS)

Feng, Jiangfan; Xu, Xin

2017-03-01

The use of geographical tagging in social-media images is becoming a part of image metadata and a great interest for geographical information science. It is well recognized that geographical topic learning is crucial for geographical annotation. Existing methods usually exploit geographical characteristics using image preprocessing, pixel-based classification, and feature recognition. How to effectively exploit the high-level semantic feature and underlying correlation among different types of contents is a crucial task for geographical topic learning. Deep learning (DL) has recently demonstrated robust capabilities for image tagging and has been introduced into geoscience. It extracts high-level features computed from a whole image component, where the cluttered background may dominate spatial features in the deep representation. Therefore, a method of spatial-attentional DL for geographical topic learning is provided and we can regard it as a special case of DL combined with various deep networks and tuning tricks. Results demonstrated that the method is discriminative for different types of geographical topic learning. In addition, it outperforms other sequential processing models in a tagging task for a geographical image dataset.

Local curvature analysis for classifying breast tumors: Preliminary analysis in dedicated breast CT

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

Lee, Juhun, E-mail: leej15@upmc.edu; Nishikawa, Robert M.; Reiser, Ingrid

2015-09-15

Purpose: The purpose of this study is to measure the effectiveness of local curvature measures as novel image features for classifying breast tumors. Methods: A total of 119 breast lesions from 104 noncontrast dedicated breast computed tomography images of women were used in this study. Volumetric segmentation was done using a seed-based segmentation algorithm and then a triangulated surface was extracted from the resulting segmentation. Total, mean, and Gaussian curvatures were then computed. Normalized curvatures were used as classification features. In addition, traditional image features were also extracted and a forward feature selection scheme was used to select the optimalmore » feature set. Logistic regression was used as a classifier and leave-one-out cross-validation was utilized to evaluate the classification performances of the features. The area under the receiver operating characteristic curve (AUC, area under curve) was used as a figure of merit. Results: Among curvature measures, the normalized total curvature (C{sub T}) showed the best classification performance (AUC of 0.74), while the others showed no classification power individually. Five traditional image features (two shape, two margin, and one texture descriptors) were selected via the feature selection scheme and its resulting classifier achieved an AUC of 0.83. Among those five features, the radial gradient index (RGI), which is a margin descriptor, showed the best classification performance (AUC of 0.73). A classifier combining RGI and C{sub T} yielded an AUC of 0.81, which showed similar performance (i.e., no statistically significant difference) to the classifier with the above five traditional image features. Additional comparisons in AUC values between classifiers using different combinations of traditional image features and C{sub T} were conducted. The results showed that C{sub T} was able to replace the other four image features for the classification task. Conclusions: The normalized curvature measure contains useful information in classifying breast tumors. Using this, one can reduce the number of features in a classifier, which may result in more robust classifiers for different datasets.« less

Medial temporal lobe contributions to short-term memory for faces.

PubMed

Race, Elizabeth; LaRocque, Karen F; Keane, Margaret M; Verfaellie, Mieke

2013-11-01

The role of the medial temporal lobes (MTL) in short-term memory (STM) remains a matter of debate. Whereas imaging studies commonly show hippocampal activation during short-delay memory tasks, evidence from amnesic patients with MTL lesions is mixed. It has been argued that apparent STM impairments in amnesia may reflect long-term memory (LTM) contributions to performance. We challenge this conclusion by demonstrating that MTL amnesic patients show impaired delayed matching-to-sample (DMS) for faces in a task that meets both a traditional delay-based and a recently proposed distractor-based criterion for classification as an STM task. In Experiment 1, we demonstrate that our face DMS task meets the proposed distractor-based criterion for STM classification, in that extensive processing of delay-period distractor stimuli disrupts performance of healthy individuals. In Experiment 2, MTL amnesic patients with lesions extending into anterior subhippocampal cortex, but not patients with lesions limited to the hippocampus, show impaired performance on this task without distraction at delays as short as 8 s, within temporal range of delay-based STM classification, in the context of intact perceptual matching performance. Experiment 3 provides support for the hypothesis that STM for faces relies on configural processing by showing that the extent to which healthy participants' performance is disrupted by interference depends on the configural demands of the distractor task. Together, these findings are consistent with the notion that the amnesic impairment in STM for faces reflects a deficit in configural processing associated with subhippocampal cortices and provide novel evidence that the MTL supports cognition beyond the LTM domain. PsycINFO Database Record (c) 2013 APA, all rights reserved.

Cognitive approaches for patterns analysis and security applications

NASA Astrophysics Data System (ADS)

Ogiela, Marek R.; Ogiela, Lidia

2017-08-01

In this paper will be presented new opportunities for developing innovative solutions for semantic pattern classification and visual cryptography, which will base on cognitive and bio-inspired approaches. Such techniques can be used for evaluation of the meaning of analyzed patterns or encrypted information, and allow to involve such meaning into the classification task or encryption process. It also allows using some crypto-biometric solutions to extend personalized cryptography methodologies based on visual pattern analysis. In particular application of cognitive information systems for semantic analysis of different patterns will be presented, and also a novel application of such systems for visual secret sharing will be described. Visual shares for divided information can be created based on threshold procedure, which may be dependent on personal abilities to recognize some image details visible on divided images.

Finding Edges and Lines in Images.

DTIC Science & Technology

1983-06-01

34 UNCLASSI FlED , SECURITY CLASSIFICATION OF THIS PAGE ("osen Data Entered) READ INSTRUCTIONSREPORT DOCUMENTATION PAGE BEFORE COMPLETING FORM I. REPORT...PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT. PROJECT. TASK Artificial Intelligence Laboratory AREA&WORKUNITNUMBERS 545 Technology Square...in the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology. Support for the laboratory’s artificial intelligence research

Human Factors Engineering. Student Supplement,

DTIC Science & Technology

1981-08-01

a job TASK TAXONOMY A classification scheme for the different levels of activities in a system, i.e., job - task - sub-task, etc. TASK-AN~ALYSIS...with the classification of learning objectives by learning category so as to identify learningPhas III guidelines necessary for optimum learning to...correct. .4... .the sequencing of all dependent tasks. .1.. .the classification of learning objectives by learning category and the Identification of

Spectral-Spatial Shared Linear Regression for Hyperspectral Image Classification.

PubMed

Haoliang Yuan; Yuan Yan Tang

2017-04-01

Classification of the pixels in hyperspectral image (HSI) is an important task and has been popularly applied in many practical applications. Its major challenge is the high-dimensional small-sized problem. To deal with this problem, lots of subspace learning (SL) methods are developed to reduce the dimension of the pixels while preserving the important discriminant information. Motivated by ridge linear regression (RLR) framework for SL, we propose a spectral-spatial shared linear regression method (SSSLR) for extracting the feature representation. Comparing with RLR, our proposed SSSLR has the following two advantages. First, we utilize a convex set to explore the spatial structure for computing the linear projection matrix. Second, we utilize a shared structure learning model, which is formed by original data space and a hidden feature space, to learn a more discriminant linear projection matrix for classification. To optimize our proposed method, an efficient iterative algorithm is proposed. Experimental results on two popular HSI data sets, i.e., Indian Pines and Salinas demonstrate that our proposed methods outperform many SL methods.

An eye tracking study of bloodstain pattern analysts during pattern classification.

PubMed

Arthur, R M; Hoogenboom, J; Green, R D; Taylor, M C; de Bruin, K G

2018-05-01

Bloodstain pattern analysis (BPA) is the forensic discipline concerned with the classification and interpretation of bloodstains and bloodstain patterns at the crime scene. At present, it is unclear exactly which stain or pattern properties and their associated values are most relevant to analysts when classifying a bloodstain pattern. Eye tracking technology has been widely used to investigate human perception and cognition. Its application to forensics, however, is limited. This is the first study to use eye tracking as a tool for gaining access to the mindset of the bloodstain pattern expert. An eye tracking method was used to follow the gaze of 24 bloodstain pattern analysts during an assigned task of classifying a laboratory-generated test bloodstain pattern. With the aid of an automated image-processing methodology, the properties of selected features of the pattern were quantified leading to the delineation of areas of interest (AOIs). Eye tracking data were collected for each AOI and combined with verbal statements made by analysts after the classification task to determine the critical range of values for relevant diagnostic features. Eye-tracking data indicated that there were four main regions of the pattern that analysts were most interested in. Within each region, individual elements or groups of elements that exhibited features associated with directionality, size, colour and shape appeared to capture the most interest of analysts during the classification task. The study showed that the eye movements of trained bloodstain pattern experts and their verbal descriptions of a pattern were well correlated.

Voting for image scoring and assessment (VISA)--theory and application of a 2 + 1 reader algorithm to improve accuracy of imaging endpoints in clinical trials.

PubMed

Gottlieb, Klaus; Hussain, Fez

2015-02-19

Independent central reading or off-site reading of imaging endpoints is increasingly used in clinical trials. Clinician-reported outcomes, such as endoscopic disease activity scores, have been shown to be subject to bias and random error. Central reading attempts to limit bias and improve accuracy of the assessment, two factors that are critical to trial success. Whether one central reader is sufficient and how to best integrate the input of more than one central reader into one output measure, is currently not known.In this concept paper we develop the theoretical foundations of a reading algorithm that can achieve both objectives without jeopardizing operational efficiency We examine the role of expert versus competent reader, frame scoring of imaging as a classification task, and propose a voting algorithm (VISA: Voting for Image Scoring and Assessment) as the most appropriate solution which could also be used to operationally define imaging gold standards. We propose two image readers plus an optional third reader in cases of disagreement (2 + 1) for ordinary scoring tasks. We argue that it is critical in trials with endoscopically determined endpoints to include the score determined by the site reader, at least in endoscopy clinical trials. Juries with more than 3 readers could define a reference standard that would allow a transition from measuring reader agreement to measuring reader accuracy. We support VISA by applying concepts from engineering (triple-modular redundancy) and voting theory (Condorcet's jury theorem) and illustrate our points with examples from inflammatory bowel disease trials, specifically, the endoscopy component of the Mayo Clinic Score of ulcerative colitis disease activity. Detailed flow-diagrams (pseudo-code) are provided that can inform program design.The VISA "2 + 1" reading algorithm, based on voting, can translate individual reader scores into a final score in a fashion that is both mathematically sound (by avoiding averaging of ordinal data) and in a manner that is consistent with the scoring task at hand (based on decisions about the presence or absence of features, a subjective classification task). While the VISA 2 + 1 algorithm is currently being used in clinical trials, empirical data of its performance have not yet been reported.

Fast parallel image registration on CPU and GPU for diagnostic classification of Alzheimer's disease

PubMed Central

Shamonin, Denis P.; Bron, Esther E.; Lelieveldt, Boudewijn P. F.; Smits, Marion; Klein, Stefan; Staring, Marius

2013-01-01

Nonrigid image registration is an important, but time-consuming task in medical image analysis. In typical neuroimaging studies, multiple image registrations are performed, i.e., for atlas-based segmentation or template construction. Faster image registration routines would therefore be beneficial. In this paper we explore acceleration of the image registration package elastix by a combination of several techniques: (i) parallelization on the CPU, to speed up the cost function derivative calculation; (ii) parallelization on the GPU building on and extending the OpenCL framework from ITKv4, to speed up the Gaussian pyramid computation and the image resampling step; (iii) exploitation of certain properties of the B-spline transformation model; (iv) further software optimizations. The accelerated registration tool is employed in a study on diagnostic classification of Alzheimer's disease and cognitively normal controls based on T1-weighted MRI. We selected 299 participants from the publicly available Alzheimer's Disease Neuroimaging Initiative database. Classification is performed with a support vector machine based on gray matter volumes as a marker for atrophy. We evaluated two types of strategies (voxel-wise and region-wise) that heavily rely on nonrigid image registration. Parallelization and optimization resulted in an acceleration factor of 4–5x on an 8-core machine. Using OpenCL a speedup factor of 2 was realized for computation of the Gaussian pyramids, and 15–60 for the resampling step, for larger images. The voxel-wise and the region-wise classification methods had an area under the receiver operator characteristic curve of 88 and 90%, respectively, both for standard and accelerated registration. We conclude that the image registration package elastix was substantially accelerated, with nearly identical results to the non-optimized version. The new functionality will become available in the next release of elastix as open source under the BSD license. PMID:24474917

Cross-domain and multi-task transfer learning of deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis

NASA Astrophysics Data System (ADS)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir; Helvie, Mark A.; Richter, Caleb; Cha, Kenny

2018-02-01

We propose a cross-domain, multi-task transfer learning framework to transfer knowledge learned from non-medical images by a deep convolutional neural network (DCNN) to medical image recognition task while improving the generalization by multi-task learning of auxiliary tasks. A first stage cross-domain transfer learning was initiated from ImageNet trained DCNN to mammography trained DCNN. 19,632 regions-of-interest (ROI) from 2,454 mass lesions were collected from two imaging modalities: digitized-screen film mammography (SFM) and full-field digital mammography (DM), and split into training and test sets. In the multi-task transfer learning, the DCNN learned the mass classification task simultaneously from the training set of SFM and DM. The best transfer network for mammography was selected from three transfer networks with different number of convolutional layers frozen. The performance of single-task and multitask transfer learning on an independent SFM test set in terms of the area under the receiver operating characteristic curve (AUC) was 0.78+/-0.02 and 0.82+/-0.02, respectively. In the second stage cross-domain transfer learning, a set of 12,680 ROIs from 317 mass lesions on DBT were split into validation and independent test sets. We first studied the data requirements for the first stage mammography trained DCNN by varying the mammography training data from 1% to 100% and evaluated its learning on the DBT validation set in inference mode. We found that the entire available mammography set provided the best generalization. The DBT validation set was then used to train only the last four fully connected layers, resulting in an AUC of 0.90+/-0.04 on the independent DBT test set.

Decoding of Ankle Flexion and Extension from Cortical Current Sources Estimated from Non-invasive Brain Activity Recording Methods.

PubMed

Mejia Tobar, Alejandra; Hyoudou, Rikiya; Kita, Kahori; Nakamura, Tatsuhiro; Kambara, Hiroyuki; Ogata, Yousuke; Hanakawa, Takashi; Koike, Yasuharu; Yoshimura, Natsue

2017-01-01

The classification of ankle movements from non-invasive brain recordings can be applied to a brain-computer interface (BCI) to control exoskeletons, prosthesis, and functional electrical stimulators for the benefit of patients with walking impairments. In this research, ankle flexion and extension tasks at two force levels in both legs, were classified from cortical current sources estimated by a hierarchical variational Bayesian method, using electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) recordings. The hierarchical prior for the current source estimation from EEG was obtained from activated brain areas and their intensities from an fMRI group (second-level) analysis. The fMRI group analysis was performed on regions of interest defined over the primary motor cortex, the supplementary motor area, and the somatosensory area, which are well-known to contribute to movement control. A sparse logistic regression method was applied for a nine-class classification (eight active tasks and a resting control task) obtaining a mean accuracy of 65.64% for time series of current sources, estimated from the EEG and the fMRI signals using a variational Bayesian method, and a mean accuracy of 22.19% for the classification of the pre-processed of EEG sensor signals, with a chance level of 11.11%. The higher classification accuracy of current sources, when compared to EEG classification accuracy, was attributed to the high number of sources and the different signal patterns obtained in the same vertex for different motor tasks. Since the inverse filter estimation for current sources can be done offline with the present method, the present method is applicable to real-time BCIs. Finally, due to the highly enhanced spatial distribution of current sources over the brain cortex, this method has the potential to identify activation patterns to design BCIs for the control of an affected limb in patients with stroke, or BCIs from motor imagery in patients with spinal cord injury.

Recognizing Banknote Fitness with a Visible Light One Dimensional Line Image Sensor

PubMed Central

Pham, Tuyen Danh; Park, Young Ho; Kwon, Seung Yong; Nguyen, Dat Tien; Vokhidov, Husan; Park, Kang Ryoung; Jeong, Dae Sik; Yoon, Sungsoo

2015-01-01

In general, dirty banknotes that have creases or soiled surfaces should be replaced by new banknotes, whereas clean banknotes should be recirculated. Therefore, the accurate classification of banknote fitness when sorting paper currency is an important and challenging task. Most previous research has focused on sensors that used visible, infrared, and ultraviolet light. Furthermore, there was little previous research on the fitness classification for Indian paper currency. Therefore, we propose a new method for classifying the fitness of Indian banknotes, with a one-dimensional line image sensor that uses only visible light. The fitness of banknotes is usually determined by various factors such as soiling, creases, and tears, etc. although we just consider banknote soiling in our research. This research is novel in the following four ways: first, there has been little research conducted on fitness classification for the Indian Rupee using visible-light images. Second, the classification is conducted based on the features extracted from the regions of interest (ROIs), which contain little texture. Third, 1-level discrete wavelet transformation (DWT) is used to extract the features for discriminating between fit and unfit banknotes. Fourth, the optimal DWT features that represent the fitness and unfitness of banknotes are selected based on linear regression analysis with ground-truth data measured by densitometer. In addition, the selected features are used as the inputs to a support vector machine (SVM) for the final classification of banknote fitness. Experimental results showed that our method outperforms other methods. PMID:26343654

Development of algorithms for detecting citrus canker based on hyperspectral reflectance imaging.

PubMed

Li, Jiangbo; Rao, Xiuqin; Ying, Yibin

2012-01-15

Automated discrimination of fruits with canker from other fruit with normal surface and different type of peel defects has become a helpful task to enhance the competitiveness and profitability of the citrus industry. Over the last several years, hyperspectral imaging technology has received increasing attention in the agricultural products inspection field. This paper studied the feasibility of classification of citrus canker from other peel conditions including normal surface and nine peel defects by hyperspectal imaging. A combination algorithm based on principal component analysis and the two-band ratio (Q(687/630)) method was proposed. Since fewer wavelengths were desired in order to develop a rapid multispectral imaging system, the canker classification performance of the two-band ratio (Q(687/630)) method alone was also evaluated. The proposed combination approach and two-band ratio method alone resulted in overall classification accuracy for training set samples and test set samples of 99.5%, 84.5% and 98.2%, 82.9%, respectively. The proposed combination approach was more efficient for classifying canker against various conditions under reflectance hyperspectral imagery. However, the two-band ratio (Q(687/630)) method alone also demonstrated effectiveness in discriminating citrus canker from normal fruit and other peel diseases except for copper burn and anthracnose. Copyright © 2011 Society of Chemical Industry.

Bag-of-features approach for improvement of lung tissue classification in diffuse lung disease

NASA Astrophysics Data System (ADS)

Kato, Noriji; Fukui, Motofumi; Isozaki, Takashi

2009-02-01

Many automated techniques have been proposed to classify diffuse lung disease patterns. Most of the techniques utilize texture analysis approaches with second and higher order statistics, and show successful classification result among various lung tissue patterns. However, the approaches do not work well for the patterns with inhomogeneous texture distribution within a region of interest (ROI), such as reticular and honeycombing patterns, because the statistics can only capture averaged feature over the ROI. In this work, we have introduced the bag-of-features approach to overcome this difficulty. In the approach, texture images are represented as histograms or distributions of a few basic primitives, which are obtained by clustering local image features. The intensity descriptor and the Scale Invariant Feature Transformation (SIFT) descriptor are utilized to extract the local features, which have significant discriminatory power due to their specificity to a particular image class. In contrast, the drawback of the local features is lack of invariance under translation and rotation. We improved the invariance by sampling many local regions so that the distribution of the local features is unchanged. We evaluated the performance of our system in the classification task with 5 image classes (ground glass, reticular, honeycombing, emphysema, and normal) using 1109 ROIs from 211 patients. Our system achieved high classification accuracy of 92.8%, which is superior to that of the conventional system with the gray level co-occurrence matrix (GLCM) feature especially for inhomogeneous texture patterns.

A Robust Deep Model for Improved Classification of AD/MCI Patients

PubMed Central

Li, Feng; Tran, Loc; Thung, Kim-Han; Ji, Shuiwang; Shen, Dinggang; Li, Jiang

2015-01-01

Accurate classification of Alzheimer’s Disease (AD) and its prodromal stage, Mild Cognitive Impairment (MCI), plays a critical role in possibly preventing progression of memory impairment and improving quality of life for AD patients. Among many research tasks, it is of particular interest to identify noninvasive imaging biomarkers for AD diagnosis. In this paper, we present a robust deep learning system to identify different progression stages of AD patients based on MRI and PET scans. We utilized the dropout technique to improve classical deep learning by preventing its weight co-adaptation, which is a typical cause of over-fitting in deep learning. In addition, we incorporated stability selection, an adaptive learning factor, and a multi-task learning strategy into the deep learning framework. We applied the proposed method to the ADNI data set and conducted experiments for AD and MCI conversion diagnosis. Experimental results showed that the dropout technique is very effective in AD diagnosis, improving the classification accuracies by 5.9% on average as compared to the classical deep learning methods. PMID:25955998

Subliminal priming with nearly perfect performance in the prime-classification task.

PubMed

Finkbeiner, Matthew

2011-05-01

The subliminal priming paradigm is widely used by cognitive scientists, and claims of subliminal perception are common nowadays. Nevertheless, there are still those who remain skeptical. In a recent critique of subliminal priming, Pratte and Rouder (Attention, Perception, & Psychophysics, 71, 1276-1283, 2009) suggested that previous claims of subliminal priming may have been due to a failure to control the task difficulty between the experiment proper and the prime-classification task. Essentially, because the prime-classification task is more difficult than the experiment proper, the prime-classification task results may underrepresent the subjects' true ability to perceive the prime stimuli. To address this possibility, prime words were here presented in color. In the experiment proper, priming was observed. In the prime-classification task, subjects reported the color of the primes very accurately, indicating almost perfect control of task difficulty, but they could not identify the primes. Thus, I conclude that controlling for task difficulty does not eliminate subliminal priming.

Assessing microscope image focus quality with deep learning.

PubMed

Yang, Samuel J; Berndl, Marc; Michael Ando, D; Barch, Mariya; Narayanaswamy, Arunachalam; Christiansen, Eric; Hoyer, Stephan; Roat, Chris; Hung, Jane; Rueden, Curtis T; Shankar, Asim; Finkbeiner, Steven; Nelson, Philip

2018-03-15

Large image datasets acquired on automated microscopes typically have some fraction of low quality, out-of-focus images, despite the use of hardware autofocus systems. Identification of these images using automated image analysis with high accuracy is important for obtaining a clean, unbiased image dataset. Complicating this task is the fact that image focus quality is only well-defined in foreground regions of images, and as a result, most previous approaches only enable a computation of the relative difference in quality between two or more images, rather than an absolute measure of quality. We present a deep neural network model capable of predicting an absolute measure of image focus on a single image in isolation, without any user-specified parameters. The model operates at the image-patch level, and also outputs a measure of prediction certainty, enabling interpretable predictions. The model was trained on only 384 in-focus Hoechst (nuclei) stain images of U2OS cells, which were synthetically defocused to one of 11 absolute defocus levels during training. The trained model can generalize on previously unseen real Hoechst stain images, identifying the absolute image focus to within one defocus level (approximately 3 pixel blur diameter difference) with 95% accuracy. On a simpler binary in/out-of-focus classification task, the trained model outperforms previous approaches on both Hoechst and Phalloidin (actin) stain images (F-scores of 0.89 and 0.86, respectively over 0.84 and 0.83), despite only having been presented Hoechst stain images during training. Lastly, we observe qualitatively that the model generalizes to two additional stains, Hoechst and Tubulin, of an unseen cell type (Human MCF-7) acquired on a different instrument. Our deep neural network enables classification of out-of-focus microscope images with both higher accuracy and greater precision than previous approaches via interpretable patch-level focus and certainty predictions. The use of synthetically defocused images precludes the need for a manually annotated training dataset. The model also generalizes to different image and cell types. The framework for model training and image prediction is available as a free software library and the pre-trained model is available for immediate use in Fiji (ImageJ) and CellProfiler.

Face recognition: database acquisition, hybrid algorithms, and human studies

NASA Astrophysics Data System (ADS)

Gutta, Srinivas; Huang, Jeffrey R.; Singh, Dig; Wechsler, Harry

1997-02-01

One of the most important technologies absent in traditional and emerging frontiers of computing is the management of visual information. Faces are accessible `windows' into the mechanisms that govern our emotional and social lives. The corresponding face recognition tasks considered herein include: (1) Surveillance, (2) CBIR, and (3) CBIR subject to correct ID (`match') displaying specific facial landmarks such as wearing glasses. We developed robust matching (`classification') and retrieval schemes based on hybrid classifiers and showed their feasibility using the FERET database. The hybrid classifier architecture consist of an ensemble of connectionist networks--radial basis functions-- and decision trees. The specific characteristics of our hybrid architecture include (a) query by consensus as provided by ensembles of networks for coping with the inherent variability of the image formation and data acquisition process, and (b) flexible and adaptive thresholds as opposed to ad hoc and hard thresholds. Experimental results, proving the feasibility of our approach, yield (i) 96% accuracy, using cross validation (CV), for surveillance on a data base consisting of 904 images (ii) 97% accuracy for CBIR tasks, on a database of 1084 images, and (iii) 93% accuracy, using CV, for CBIR subject to correct ID match tasks on a data base of 200 images.

Recognition of skin melanoma through dermoscopic image analysis

NASA Astrophysics Data System (ADS)

Gómez, Catalina; Herrera, Diana Sofia

2017-11-01

Melanoma skin cancer diagnosis can be challenging due to the similarities of the early stage symptoms with regular moles. Standardized visual parameters can be determined and characterized to suspect a melanoma cancer type. The automation of this diagnosis could have an impact in the medical field by providing a tool to support the specialists with high accuracy. The objective of this study is to develop an algorithm trained to distinguish a highly probable melanoma from a non-dangerous mole by the segmentation and classification of dermoscopic mole images. We evaluate our approach on the dataset provided by the International Skin Imaging Collaboration used in the International Challenge Skin Lesion Analysis Towards Melanoma Detection. For the segmentation task, we apply a preprocessing algorithm and use Otsu's thresholding in the best performing color space; the average Jaccard Index in the test dataset is 70.05%. For the subsequent classification stage, we use joint histograms in the YCbCr color space, a RBF Gaussian SVM trained with five features concerning circularity and irregularity of the segmented lesion, and the Gray Level Co-occurrence matrix features for texture analysis. These features are combined to obtain an Average Classification Accuracy of 63.3% in the test dataset.

Constrained Metric Learning by Permutation Inducing Isometries.

PubMed

Bosveld, Joel; Mahmood, Arif; Huynh, Du Q; Noakes, Lyle

2016-01-01

The choice of metric critically affects the performance of classification and clustering algorithms. Metric learning algorithms attempt to improve performance, by learning a more appropriate metric. Unfortunately, most of the current algorithms learn a distance function which is not invariant to rigid transformations of images. Therefore, the distances between two images and their rigidly transformed pair may differ, leading to inconsistent classification or clustering results. We propose to constrain the learned metric to be invariant to the geometry preserving transformations of images that induce permutations in the feature space. The constraint that these transformations are isometries of the metric ensures consistent results and improves accuracy. Our second contribution is a dimension reduction technique that is consistent with the isometry constraints. Our third contribution is the formulation of the isometry constrained logistic discriminant metric learning (IC-LDML) algorithm, by incorporating the isometry constraints within the objective function of the LDML algorithm. The proposed algorithm is compared with the existing techniques on the publicly available labeled faces in the wild, viewpoint-invariant pedestrian recognition, and Toy Cars data sets. The IC-LDML algorithm has outperformed existing techniques for the tasks of face recognition, person identification, and object classification by a significant margin.

The footprints of visual attention in the Posner cueing paradigm revealed by classification images

NASA Technical Reports Server (NTRS)

Eckstein, Miguel P.; Shimozaki, Steven S.; Abbey, Craig K.

2002-01-01

In the Posner cueing paradigm, observers' performance in detecting a target is typically better in trials in which the target is present at the cued location than in trials in which the target appears at the uncued location. This effect can be explained in terms of a Bayesian observer where visual attention simply weights the information differently at the cued (attended) and uncued (unattended) locations without a change in the quality of processing at each location. Alternatively, it could also be explained in terms of visual attention changing the shape of the perceptual filter at the cued location. In this study, we use the classification image technique to compare the human perceptual filters at the cued and uncued locations in a contrast discrimination task. We did not find statistically significant differences between the shapes of the inferred perceptual filters across the two locations, nor did the observed differences account for the measured cueing effects in human observers. Instead, we found a difference in the magnitude of the classification images, supporting the idea that visual attention changes the weighting of information at the cued and uncued location, but does not change the quality of processing at each individual location.

Computer-aided diagnosis in phase contrast imaging X-ray computed tomography for quantitative characterization of ex vivo human patellar cartilage.

PubMed

Nagarajan, Mahesh B; Coan, Paola; Huber, Markus B; Diemoz, Paul C; Glaser, Christian; Wismuller, Axel

2013-10-01

Visualization of ex vivo human patellar cartilage matrix through the phase contrast imaging X-ray computed tomography (PCI-CT) has been previously demonstrated. Such studies revealed osteoarthritis-induced changes to chondrocyte organization in the radial zone. This study investigates the application of texture analysis to characterizing such chondrocyte patterns in the presence and absence of osteoarthritic damage. Texture features derived from Minkowski functionals (MF) and gray-level co-occurrence matrices (GLCM) were extracted from 842 regions of interest (ROI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. These texture features were subsequently used in a machine learning task with support vector regression to classify ROIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver operating characteristic curve (AUC). The best classification performance was observed with the MF features perimeter (AUC: 0.94 ±0.08 ) and "Euler characteristic" (AUC: 0.94 ±0.07 ), and GLCM-derived feature "Correlation" (AUC: 0.93 ±0.07). These results suggest that such texture features can provide a detailed characterization of the chondrocyte organization in the cartilage matrix, enabling classification of cartilage as healthy or osteoarthritic with high accuracy.

Classification of change detection and change blindness from near-infrared spectroscopy signals

NASA Astrophysics Data System (ADS)

Tanaka, Hirokazu; Katura, Takusige

2011-08-01

Using a machine-learning classification algorithm applied to near-infrared spectroscopy (NIRS) signals, we classify a success (change detection) or a failure (change blindness) in detecting visual changes for a change-detection task. Five subjects perform a change-detection task, and their brain activities are continuously monitored. A support-vector-machine algorithm is applied to classify the change-detection and change-blindness trials, and correct classification probability of 70-90% is obtained for four subjects. Two types of temporal shapes in classification probabilities are found: one exhibiting a maximum value after the task is completed (postdictive type), and another exhibiting a maximum value during the task (predictive type). As for the postdictive type, the classification probability begins to increase immediately after the task completion and reaches its maximum in about the time scale of neuronal hemodynamic response, reflecting a subjective report of change detection. As for the predictive type, the classification probability shows an increase at the task initiation and is maximal while subjects are performing the task, predicting the task performance in detecting a change. We conclude that decoding change detection and change blindness from NIRS signal is possible and argue some future applications toward brain-machine interfaces.

The use of kernel local Fisher discriminant analysis for the channelization of the Hotelling model observer

NASA Astrophysics Data System (ADS)

Wen, Gezheng; Markey, Mia K.

2015-03-01

It is resource-intensive to conduct human studies for task-based assessment of medical image quality and system optimization. Thus, numerical model observers have been developed as a surrogate for human observers. The Hotelling observer (HO) is the optimal linear observer for signal-detection tasks, but the high dimensionality of imaging data results in a heavy computational burden. Channelization is often used to approximate the HO through a dimensionality reduction step, but how to produce channelized images without losing significant image information remains a key challenge. Kernel local Fisher discriminant analysis (KLFDA) uses kernel techniques to perform supervised dimensionality reduction, which finds an embedding transformation that maximizes betweenclass separability and preserves within-class local structure in the low-dimensional manifold. It is powerful for classification tasks, especially when the distribution of a class is multimodal. Such multimodality could be observed in many practical clinical tasks. For example, primary and metastatic lesions may both appear in medical imaging studies, but the distributions of their typical characteristics (e.g., size) may be very different. In this study, we propose to use KLFDA as a novel channelization method. The dimension of the embedded manifold (i.e., the result of KLFDA) is a counterpart to the number of channels in the state-of-art linear channelization. We present a simulation study to demonstrate the potential usefulness of KLFDA for building the channelized HOs (CHOs) and generating reliable decision statistics for clinical tasks. We show that the performance of the CHO with KLFDA channels is comparable to that of the benchmark CHOs.

Benthic Habitat Mapping by Combining Lyzenga’s Optical Model and Relative Water Depth Model in Lintea Island, Southeast Sulawesi

NASA Astrophysics Data System (ADS)

Hafizt, M.; Manessa, M. D. M.; Adi, N. S.; Prayudha, B.

2017-12-01

Benthic habitat mapping using satellite data is one challenging task for practitioners and academician as benthic objects are covered by light-attenuating water column obscuring object discrimination. One common method to reduce this water-column effect is by using depth-invariant index (DII) image. However, the application of the correction in shallow coastal areas is challenging as a dark object such as seagrass could have a very low pixel value, preventing its reliable identification and classification. This limitation can be solved by specifically applying a classification process to areas with different water depth levels. The water depth level can be extracted from satellite imagery using Relative Water Depth Index (RWDI). This study proposed a new approach to improve the mapping accuracy, particularly for benthic dark objects by combining the DII of Lyzenga’s water column correction method and the RWDI of Stumpt’s method. This research was conducted in Lintea Island which has a high variation of benthic cover using Sentinel-2A imagery. To assess the effectiveness of the proposed new approach for benthic habitat mapping two different classification procedures are implemented. The first procedure is the commonly applied method in benthic habitat mapping where DII image is used as input data to all coastal area for image classification process regardless of depth variation. The second procedure is the proposed new approach where its initial step begins with the separation of the study area into shallow and deep waters using the RWDI image. Shallow area was then classified using the sunglint-corrected image as input data and the deep area was classified using DII image as input data. The final classification maps of those two areas were merged as a single benthic habitat map. A confusion matrix was then applied to evaluate the mapping accuracy of the final map. The result shows that the new proposed mapping approach can be used to map all benthic objects in all depth ranges and shows a better accuracy compared to that of classification map produced using only with DII.

Optimality of the basic colour categories for classification

PubMed Central

Griffin, Lewis D

2005-01-01

Categorization of colour has been widely studied as a window into human language and cognition, and quite separately has been used pragmatically in image-database retrieval systems. This suggests the hypothesis that the best category system for pragmatic purposes coincides with human categories (i.e. the basic colours). We have tested this hypothesis by assessing the performance of different category systems in a machine-vision task. The task was the identification of the odd-one-out from triples of images obtained using a web-based image-search service. In each triple, two of the images had been retrieved using the same search term, the other a different term. The terms were simple concrete nouns. The results were as follows: (i) the odd-one-out task can be performed better than chance using colour alone; (ii) basic colour categorization performs better than random systems of categories; (iii) a category system that performs better than the basic colours could not be found; and (iv) it is not just the general layout of the basic colours that is important, but also the detail. We conclude that (i) the results support the plausibility of an explanation for the basic colours as a result of a pressure-to-optimality and (ii) the basic colours are good categories for machine vision image-retrieval systems. PMID:16849219

Neural and Behavioral Sequelae of Blast-Related Traumatic Brain Injury

DTIC Science & Technology

2012-11-01

testing and advanced MRI techniques [task-activated functional MRI (fMRI) and diffusion tensor imaging ( DTI )] to gain a comprehensive understanding of... DTI fiber tracking) and neurobehavioral testing (computerized assessment and standard neuropsychological testing) on 60 chronic trauma patients: 15...data analysis. 15. SUBJECT TERMS Blast-related traumatic brain injury (TBI), fMRI, DTI , cognition 16. SECURITY CLASSIFICATION OF: 17. LIMITATION

Clusters of Word Properties as Predictors of Elementary School Children's Performance on Two Word Tasks

ERIC Educational Resources Information Center

Tellings, Agnes; Coppens, Karien; Gelissen, John; Schreuder, Rob

2013-01-01

Often, the classification of words does not go beyond "difficult" (i.e., infrequent, late-learned, nonimageable, etc.) or "easy" (i.e., frequent, early-learned, imageable, etc.) words. In the present study, we used a latent cluster analysis to divide 703 Dutch words with scores for eight word properties into seven clusters of words. Each cluster…

3D Texture Features Mining for MRI Brain Tumor Identification

NASA Astrophysics Data System (ADS)

Rahim, Mohd Shafry Mohd; Saba, Tanzila; Nayer, Fatima; Syed, Afraz Zahra

2014-03-01

Medical image segmentation is a process to extract region of interest and to divide an image into its individual meaningful, homogeneous components. Actually, these components will have a strong relationship with the objects of interest in an image. For computer-aided diagnosis and therapy process, medical image segmentation is an initial mandatory step. Medical image segmentation is a sophisticated and challenging task because of the sophisticated nature of the medical images. Indeed, successful medical image analysis heavily dependent on the segmentation accuracy. Texture is one of the major features to identify region of interests in an image or to classify an object. 2D textures features yields poor classification results. Hence, this paper represents 3D features extraction using texture analysis and SVM as segmentation technique in the testing methodologies.

Data analysis in emission tomography using emission-count posteriors

NASA Astrophysics Data System (ADS)

Sitek, Arkadiusz

2012-11-01

A novel approach to the analysis of emission tomography data using the posterior probability of the number of emissions per voxel (emission count) conditioned on acquired tomographic data is explored. The posterior is derived from the prior and the Poisson likelihood of the emission-count data by marginalizing voxel activities. Based on emission-count posteriors, examples of Bayesian analysis including estimation and classification tasks in emission tomography are provided. The application of the method to computer simulations of 2D tomography is demonstrated. In particular, the minimum-mean-square-error point estimator of the emission count is demonstrated. The process of finding this estimator can be considered as a tomographic image reconstruction technique since the estimates of the number of emissions per voxel divided by voxel sensitivities and acquisition time are the estimates of the voxel activities. As an example of a classification task, a hypothesis stating that some region of interest (ROI) emitted at least or at most r-times the number of events in some other ROI is tested. The ROIs are specified by the user. The analysis described in this work provides new quantitative statistical measures that can be used in decision making in diagnostic imaging using emission tomography.

Constrained Low-Rank Learning Using Least Squares-Based Regularization.

PubMed

Li, Ping; Yu, Jun; Wang, Meng; Zhang, Luming; Cai, Deng; Li, Xuelong

2017-12-01

Low-rank learning has attracted much attention recently due to its efficacy in a rich variety of real-world tasks, e.g., subspace segmentation and image categorization. Most low-rank methods are incapable of capturing low-dimensional subspace for supervised learning tasks, e.g., classification and regression. This paper aims to learn both the discriminant low-rank representation (LRR) and the robust projecting subspace in a supervised manner. To achieve this goal, we cast the problem into a constrained rank minimization framework by adopting the least squares regularization. Naturally, the data label structure tends to resemble that of the corresponding low-dimensional representation, which is derived from the robust subspace projection of clean data by low-rank learning. Moreover, the low-dimensional representation of original data can be paired with some informative structure by imposing an appropriate constraint, e.g., Laplacian regularizer. Therefore, we propose a novel constrained LRR method. The objective function is formulated as a constrained nuclear norm minimization problem, which can be solved by the inexact augmented Lagrange multiplier algorithm. Extensive experiments on image classification, human pose estimation, and robust face recovery have confirmed the superiority of our method.

2D image classification for 3D anatomy localization: employing deep convolutional neural networks

NASA Astrophysics Data System (ADS)

de Vos, Bob D.; Wolterink, Jelmer M.; de Jong, Pim A.; Viergever, Max A.; Išgum, Ivana

2016-03-01

Localization of anatomical regions of interest (ROIs) is a preprocessing step in many medical image analysis tasks. While trivial for humans, it is complex for automatic methods. Classic machine learning approaches require the challenge of hand crafting features to describe differences between ROIs and background. Deep convolutional neural networks (CNNs) alleviate this by automatically finding hierarchical feature representations from raw images. We employ this trait to detect anatomical ROIs in 2D image slices in order to localize them in 3D. In 100 low-dose non-contrast enhanced non-ECG synchronized screening chest CT scans, a reference standard was defined by manually delineating rectangular bounding boxes around three anatomical ROIs -- heart, aortic arch, and descending aorta. Every anatomical ROI was automatically identified using a combination of three CNNs, each analyzing one orthogonal image plane. While single CNNs predicted presence or absence of a specific ROI in the given plane, the combination of their results provided a 3D bounding box around it. Classification performance of each CNN, expressed in area under the receiver operating characteristic curve, was >=0.988. Additionally, the performance of ROI localization was evaluated. Median Dice scores for automatically determined bounding boxes around the heart, aortic arch, and descending aorta were 0.89, 0.70, and 0.85 respectively. The results demonstrate that accurate automatic 3D localization of anatomical structures by CNN-based 2D image classification is feasible.

Classification of foods by transferring knowledge from ImageNet dataset

NASA Astrophysics Data System (ADS)

Heravi, Elnaz J.; Aghdam, Hamed H.; Puig, Domenec

2017-03-01

Automatic classification of foods is a way to control food intake and tackle with obesity. However, it is a challenging problem since foods are highly deformable and complex objects. Results on ImageNet dataset have revealed that Convolutional Neural Network has a great expressive power to model natural objects. Nonetheless, it is not trivial to train a ConvNet from scratch for classification of foods. This is due to the fact that ConvNets require large datasets and to our knowledge there is not a large public dataset of food for this purpose. Alternative solution is to transfer knowledge from trained ConvNets to the domain of foods. In this work, we study how transferable are state-of-art ConvNets to the task of food classification. We also propose a method for transferring knowledge from a bigger ConvNet to a smaller ConvNet by keeping its accuracy similar to the bigger ConvNet. Our experiments on UECFood256 datasets show that Googlenet, VGG and residual networks produce comparable results if we start transferring knowledge from appropriate layer. In addition, we show that our method is able to effectively transfer knowledge to the smaller ConvNet using unlabeled samples.

Texture and color features for tile classification

NASA Astrophysics Data System (ADS)

Baldrich, Ramon; Vanrell, Maria; Villanueva, Juan J.

1999-09-01

In this paper we present the results of a preliminary computer vision system to classify the production of a ceramic tile industry. We focus on the classification of a specific type of tiles whose production can be affected by external factors, such as humidity, temperature, origin of clays and pigments. Variations on these uncontrolled factors provoke small differences in the color and the texture of the tiles that force to classify all the production. A constant and non- subjective classification would allow avoiding devolution from customers and unnecessary stock fragmentation. The aim of this work is to simulate the human behavior on this classification task by extracting a set of features from tile images. These features are induced by definitions from experts. To compute them we need to mix color and texture information and to define global and local measures. In this work, we do not seek a general texture-color representation, we only deal with textures formed by non-oriented colored-blobs randomly distributed. New samples are classified using Discriminant Analysis functions derived from known class tile samples. The last part of the paper is devoted to explain the correction of acquired images in order to avoid time and geometry illumination changes.

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.

SonoNet: Real-Time Detection and Localisation of Fetal Standard Scan Planes in Freehand Ultrasound.

PubMed

Baumgartner, Christian F; Kamnitsas, Konstantinos; Matthew, Jacqueline; Fletcher, Tara P; Smith, Sandra; Koch, Lisa M; Kainz, Bernhard; Rueckert, Daniel

2017-11-01

Identifying and interpreting fetal standard scan planes during 2-D ultrasound mid-pregnancy examinations are highly complex tasks, which require years of training. Apart from guiding the probe to the correct location, it can be equally difficult for a non-expert to identify relevant structures within the image. Automatic image processing can provide tools to help experienced as well as inexperienced operators with these tasks. In this paper, we propose a novel method based on convolutional neural networks, which can automatically detect 13 fetal standard views in freehand 2-D ultrasound data as well as provide a localization of the fetal structures via a bounding box. An important contribution is that the network learns to localize the target anatomy using weak supervision based on image-level labels only. The network architecture is designed to operate in real-time while providing optimal output for the localization task. We present results for real-time annotation, retrospective frame retrieval from saved videos, and localization on a very large and challenging dataset consisting of images and video recordings of full clinical anomaly screenings. We found that the proposed method achieved an average F1-score of 0.798 in a realistic classification experiment modeling real-time detection, and obtained a 90.09% accuracy for retrospective frame retrieval. Moreover, an accuracy of 77.8% was achieved on the localization task.

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.

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.

Resting State fMRI Functional Connectivity-Based Classification Using a Convolutional Neural Network Architecture

PubMed Central

Meszlényi, Regina J.; Buza, Krisztian; Vidnyánszky, Zoltán

2017-01-01

Machine learning techniques have become increasingly popular in the field of resting state fMRI (functional magnetic resonance imaging) network based classification. However, the application of convolutional networks has been proposed only very recently and has remained largely unexplored. In this paper we describe a convolutional neural network architecture for functional connectome classification called connectome-convolutional neural network (CCNN). Our results on simulated datasets and a publicly available dataset for amnestic mild cognitive impairment classification demonstrate that our CCNN model can efficiently distinguish between subject groups. We also show that the connectome-convolutional network is capable to combine information from diverse functional connectivity metrics and that models using a combination of different connectivity descriptors are able to outperform classifiers using only one metric. From this flexibility follows that our proposed CCNN model can be easily adapted to a wide range of connectome based classification or regression tasks, by varying which connectivity descriptor combinations are used to train the network. PMID:29089883

A hybrid method for classifying cognitive states from fMRI data.

PubMed

Parida, S; Dehuri, S; Cho, S-B; Cacha, L A; Poznanski, R R

2015-09-01

Functional magnetic resonance imaging (fMRI) makes it possible to detect brain activities in order to elucidate cognitive-states. The complex nature of fMRI data requires under-standing of the analyses applied to produce possible avenues for developing models of cognitive state classification and improving brain activity prediction. While many models of classification task of fMRI data analysis have been developed, in this paper, we present a novel hybrid technique through combining the best attributes of genetic algorithms (GAs) and ensemble decision tree technique that consistently outperforms all other methods which are being used for cognitive-state classification. Specifically, this paper illustrates the combined effort of decision-trees ensemble and GAs for feature selection through an extensive simulation study and discusses the classification performance with respect to fMRI data. We have shown that our proposed method exhibits significant reduction of the number of features with clear edge classification accuracy over ensemble of decision-trees.

Resting State fMRI Functional Connectivity-Based Classification Using a Convolutional Neural Network Architecture.

PubMed

Meszlényi, Regina J; Buza, Krisztian; Vidnyánszky, Zoltán

2017-01-01

Machine learning techniques have become increasingly popular in the field of resting state fMRI (functional magnetic resonance imaging) network based classification. However, the application of convolutional networks has been proposed only very recently and has remained largely unexplored. In this paper we describe a convolutional neural network architecture for functional connectome classification called connectome-convolutional neural network (CCNN). Our results on simulated datasets and a publicly available dataset for amnestic mild cognitive impairment classification demonstrate that our CCNN model can efficiently distinguish between subject groups. We also show that the connectome-convolutional network is capable to combine information from diverse functional connectivity metrics and that models using a combination of different connectivity descriptors are able to outperform classifiers using only one metric. From this flexibility follows that our proposed CCNN model can be easily adapted to a wide range of connectome based classification or regression tasks, by varying which connectivity descriptor combinations are used to train the network.

Multivariate pattern analysis of fMRI data reveals deficits in distributed representations in schizophrenia

PubMed Central

Yoon, Jong H.; Tamir, Diana; Minzenberg, Michael J.; Ragland, J. Daniel; Ursu, Stefan; Carter, Cameron S.

2009-01-01

Background Multivariate pattern analysis is an alternative method of analyzing fMRI data, which is capable of decoding distributed neural representations. We applied this method to test the hypothesis of the impairment in distributed representations in schizophrenia. We also compared the results of this method with traditional GLM-based univariate analysis. Methods 19 schizophrenia and 15 control subjects viewed two runs of stimuli--exemplars of faces, scenes, objects, and scrambled images. To verify engagement with stimuli, subjects completed a 1-back matching task. A multi-voxel pattern classifier was trained to identify category-specific activity patterns on one run of fMRI data. Classification testing was conducted on the remaining run. Correlation of voxel-wise activity across runs evaluated variance over time in activity patterns. Results Patients performed the task less accurately. This group difference was reflected in the pattern analysis results with diminished classification accuracy in patients compared to controls, 59% and 72% respectively. In contrast, there was no group difference in GLM-based univariate measures. In both groups, classification accuracy was significantly correlated with behavioral measures. Both groups showed highly significant correlation between inter-run correlations and classification accuracy. Conclusions Distributed representations of visual objects are impaired in schizophrenia. This impairment is correlated with diminished task performance, suggesting that decreased integrity of cortical activity patterns is reflected in impaired behavior. Comparisons with univariate results suggest greater sensitivity of pattern analysis in detecting group differences in neural activity and reduced likelihood of non-specific factors driving these results. PMID:18822407

A survey on deep learning in medical image analysis.

PubMed

Litjens, Geert; Kooi, Thijs; Bejnordi, Babak Ehteshami; Setio, Arnaud Arindra Adiyoso; Ciompi, Francesco; Ghafoorian, Mohsen; van der Laak, Jeroen A W M; van Ginneken, Bram; Sánchez, Clara I

2017-12-01

Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images. This paper reviews the major deep learning concepts pertinent to medical image analysis and summarizes over 300 contributions to the field, most of which appeared in the last year. We survey the use of deep learning for image classification, object detection, segmentation, registration, and other tasks. Concise overviews are provided of studies per application area: neuro, retinal, pulmonary, digital pathology, breast, cardiac, abdominal, musculoskeletal. We end with a summary of the current state-of-the-art, a critical discussion of open challenges and directions for future research. Copyright © 2017 Elsevier B.V. All rights reserved.

Nonlinear filtering for character recognition in low quality document images

NASA Astrophysics Data System (ADS)

Diaz-Escobar, Julia; Kober, Vitaly

2014-09-01

Optical character recognition in scanned printed documents is a well-studied task, where the captured conditions like sheet position, illumination, contrast and resolution are controlled. Nowadays, it is more practical to use mobile devices for document capture than a scanner. So as a consequence, the quality of document images is often poor owing to presence of geometric distortions, nonhomogeneous illumination, low resolution, etc. In this work we propose to use multiple adaptive nonlinear composite filters for detection and classification of characters. Computer simulation results obtained with the proposed system are presented and discussed.

Detection of Focal Cortical Dysplasia Lesions in MRI Using Textural Features

NASA Astrophysics Data System (ADS)

Loyek, Christian; Woermann, Friedrich G.; Nattkemper, Tim W.

Focal cortical dysplasia (FCD) is a frequent cause of medically refractory partial epilepsy. The visual identification of FCD lesions on magnetic resonance images (MRI) is a challenging task in standard radiological analysis. Quantitative image analysis which tries to assist in the diagnosis of FCD lesions is an active field of research. In this work we investigate the potential of different texture features, in order to explore to what extent they are suitable for detecting lesional tissue. As a result we can show first promising results based on segmentation and texture classification.

Combining deep residual neural network features with supervised machine learning algorithms to classify diverse food image datasets.

PubMed

McAllister, Patrick; Zheng, Huiru; Bond, Raymond; Moorhead, Anne

2018-04-01

Obesity is increasing worldwide and can cause many chronic conditions such as type-2 diabetes, heart disease, sleep apnea, and some cancers. Monitoring dietary intake through food logging is a key method to maintain a healthy lifestyle to prevent and manage obesity. Computer vision methods have been applied to food logging to automate image classification for monitoring dietary intake. In this work we applied pretrained ResNet-152 and GoogleNet convolutional neural networks (CNNs), initially trained using ImageNet Large Scale Visual Recognition Challenge (ILSVRC) dataset with MatConvNet package, to extract features from food image datasets; Food 5K, Food-11, RawFooT-DB, and Food-101. Deep features were extracted from CNNs and used to train machine learning classifiers including artificial neural network (ANN), support vector machine (SVM), Random Forest, and Naive Bayes. Results show that using ResNet-152 deep features with SVM with RBF kernel can accurately detect food items with 99.4% accuracy using Food-5K validation food image dataset and 98.8% with Food-5K evaluation dataset using ANN, SVM-RBF, and Random Forest classifiers. Trained with ResNet-152 features, ANN can achieve 91.34%, 99.28% when applied to Food-11 and RawFooT-DB food image datasets respectively and SVM with RBF kernel can achieve 64.98% with Food-101 image dataset. From this research it is clear that using deep CNN features can be used efficiently for diverse food item image classification. The work presented in this research shows that pretrained ResNet-152 features provide sufficient generalisation power when applied to a range of food image classification tasks. Copyright © 2018 Elsevier Ltd. All rights reserved.

The joint use of the tangential electric field and surface Laplacian in EEG classification.

PubMed

Carvalhaes, C G; de Barros, J Acacio; Perreau-Guimaraes, M; Suppes, P

2014-01-01

We investigate the joint use of the tangential electric field (EF) and the surface Laplacian (SL) derivation as a method to improve the classification of EEG signals. We considered five classification tasks to test the validity of such approach. In all five tasks, the joint use of the components of the EF and the SL outperformed the scalar potential. The smallest effect occurred in the classification of a mental task, wherein the average classification rate was improved by 0.5 standard deviations. The largest effect was obtained in the classification of visual stimuli and corresponded to an improvement of 2.1 standard deviations.

Fine-grained recognition of plants from images.

PubMed

Šulc, Milan; Matas, Jiří

2017-01-01

Fine-grained recognition of plants from images is a challenging computer vision task, due to the diverse appearance and complex structure of plants, high intra-class variability and small inter-class differences. We review the state-of-the-art and discuss plant recognition tasks, from identification of plants from specific plant organs to general plant recognition "in the wild". We propose texture analysis and deep learning methods for different plant recognition tasks. The methods are evaluated and compared them to the state-of-the-art. Texture analysis is only applied to images with unambiguous segmentation (bark and leaf recognition), whereas CNNs are only applied when sufficiently large datasets are available. The results provide an insight in the complexity of different plant recognition tasks. The proposed methods outperform the state-of-the-art in leaf and bark classification and achieve very competitive results in plant recognition "in the wild". The results suggest that recognition of segmented leaves is practically a solved problem, when high volumes of training data are available. The generality and higher capacity of state-of-the-art CNNs makes them suitable for plant recognition "in the wild" where the views on plant organs or plants vary significantly and the difficulty is increased by occlusions and background clutter.

Automatic target recognition and detection in infrared imagery under cluttered background

NASA Astrophysics Data System (ADS)

Gundogdu, Erhan; Koç, Aykut; Alatan, A. Aydın.

2017-10-01

Visual object classification has long been studied in visible spectrum by utilizing conventional cameras. Since the labeled images has recently increased in number, it is possible to train deep Convolutional Neural Networks (CNN) with significant amount of parameters. As the infrared (IR) sensor technology has been improved during the last two decades, labeled images extracted from IR sensors have been started to be used for object detection and recognition tasks. We address the problem of infrared object recognition and detection by exploiting 15K images from the real-field with long-wave and mid-wave IR sensors. For feature learning, a stacked denoising autoencoder is trained in this IR dataset. To recognize the objects, the trained stacked denoising autoencoder is fine-tuned according to the binary classification loss of the target object. Once the training is completed, the test samples are propagated over the network, and the probability of the test sample belonging to a class is computed. Moreover, the trained classifier is utilized in a detect-by-classification method, where the classification is performed in a set of candidate object boxes and the maximum confidence score in a particular location is accepted as the score of the detected object. To decrease the computational complexity, the detection step at every frame is avoided by running an efficient correlation filter based tracker. The detection part is performed when the tracker confidence is below a pre-defined threshold. The experiments conducted on the real field images demonstrate that the proposed detection and tracking framework presents satisfactory results for detecting tanks under cluttered background.

Response Classification Images in Vernier Acuity

NASA Technical Reports Server (NTRS)

Ahumada, Albert J., Jr.; Beard, B. L.; Ellis, Stephen R. (Technical Monitor)

1997-01-01

Orientation selective and local sign mechanisms have been proposed as the basis for vernier acuity judgments. Linear image features contributing to discrimination can be determined for a two choice task by adding external noise to the images and then averaging the noises separately for the four types of stimulus/response trials. This method is applied to a vernier acuity task with different spatial separations to compare the predictions of the two theories. Three well-practiced observers were presented around 5000 trials of a vernier stimulus consisting of two dark horizontal lines (5 min by 0.3 min) within additive low-contrast white noise. Two spatial separations were tested, abutting and a 10 min horizontal separation. The task was to determine whether the target lines were aligned or vertically offset. The noises were averaged separately for the four stimulus/response trial types (e.g., stimulus = offset, response = aligned). The sum of the two 'not aligned' images was then subtracted from the sum of the 'aligned' images to obtain an overall image. Spatially smoothed images were quantized according to expected variability in the smoothed images to allow estimation of the statistical significance of image features. The response images from the 10 min separation condition are consistent with the local sign theory, having the appearance of two linear operators measuring vertical position with opposite sign. The images from the abutting stimulus have the same appearance with the two operators closer together. The image predicted by an oriented filter model is similar, but has its greatest weight in the abutting region, while the response images fall to nonsignificance there. The response correlation image method, previously demonstrated for letter discrimination, clarifies the features used in vernier acuity.

Evaluation of various mental task combinations for near-infrared spectroscopy-based brain-computer interfaces

NASA Astrophysics Data System (ADS)

Hwang, Han-Jeong; Lim, Jeong-Hwan; Kim, Do-Won; Im, Chang-Hwan

2014-07-01

A number of recent studies have demonstrated that near-infrared spectroscopy (NIRS) is a promising neuroimaging modality for brain-computer interfaces (BCIs). So far, most NIRS-based BCI studies have focused on enhancing the accuracy of the classification of different mental tasks. In the present study, we evaluated the performances of a variety of mental task combinations in order to determine the mental task pairs that are best suited for customized NIRS-based BCIs. To this end, we recorded event-related hemodynamic responses while seven participants performed eight different mental tasks. Classification accuracies were then estimated for all possible pairs of the eight mental tasks (C=28). Based on this analysis, mental task combinations with relatively high classification accuracies frequently included the following three mental tasks: "mental multiplication," "mental rotation," and "right-hand motor imagery." Specifically, mental task combinations consisting of two of these three mental tasks showed the highest mean classification accuracies. It is expected that our results will be a useful reference to reduce the time needed for preliminary tests when discovering individual-specific mental task combinations.

Urban Land Cover/use Change Detection Using High Resolution SPOT 5 and SPOT 6 Images and Urban Atlas Nomenclature

NASA Astrophysics Data System (ADS)

Akay, S. S.; Sertel, E.

2016-06-01

Urban land cover/use changes like urbanization and urban sprawl have been impacting the urban ecosystems significantly therefore determination of urban land cover/use changes is an important task to understand trends and status of urban ecosystems, to support urban planning and to aid decision-making for urban-based projects. High resolution satellite images could be used to accurately, periodically and quickly map urban land cover/use and their changes by time. This paper aims to determine urban land cover/use changes in Gaziantep city centre between 2010 and 2105 using object based images analysis and high resolution SPOT 5 and SPOT 6 images. 2.5 m SPOT 5 image obtained in 5th of June 2010 and 1.5 m SPOT 6 image obtained in 7th of July 2015 were used in this research to precisely determine land changes in five-year period. In addition to satellite images, various ancillary data namely Normalized Difference Vegetation Index (NDVI), Difference Water Index (NDWI) maps, cadastral maps, OpenStreetMaps, road maps and Land Cover maps, were integrated into the classification process to produce high accuracy urban land cover/use maps for these two years. Both images were geometrically corrected to fulfil the 1/10,000 scale geometric accuracy. Decision tree based object oriented classification was applied to identify twenty different urban land cover/use classes defined in European Urban Atlas project. Not only satellite images and satellite image-derived indices but also different thematic maps were integrated into decision tree analysis to create rule sets for accurate mapping of each class. Rule sets of each satellite image for the object based classification involves spectral, spatial and geometric parameter to automatically produce urban map of the city centre region. Total area of each class per related year and their changes in five-year period were determined and change trend in terms of class transformation were presented. Classification accuracy assessment was conducted by creating a confusion matrix to illustrate the thematic accuracy of each class.

Applications of Laser Scattering Probes to Turbulent Diffusion Flames

DTIC Science & Technology

1983-11-01

APPLICATIONS OF LASER SCATTERING PROBES TO TURBULENT DIFFUSION FLAMES u ^ j FINAL REPORT Contract N00014-80-C-0882 Submitted to Office of...Include Security Classification) Applications of Laser Scattering Probes to Turbulent Diffusion Flames PROJECT NO. TASK NO. WORK UNIT NO. 12...for a co-flowing jet turbulent diffusion flame, and planar laser-induced fluorescence to provide two- dimensional instantaneous images of the flame

Local Histograms for Per-Pixel Classification

DTIC Science & Technology

2012-03-01

few axioms for such models are presented. These axioms are shown to be satisfied using the convergence of random wavelet expansions. The authors of...pathologists can accurately and consistently identify and delineate tissues and their pathologies , it is an expensive and time-consuming task, therefore...Automatic Identification and Delineation of Tissues and Pathologies in H&E Stained Images. PhD Thesis. Carnegie Mellon University, Pittsburgh, PA (September

A framework for biomedical figure segmentation towards image-based document retrieval

PubMed Central

2013-01-01

The figures included in many of the biomedical publications play an important role in understanding the biological experiments and facts described within. Recent studies have shown that it is possible to integrate the information that is extracted from figures in classical document classification and retrieval tasks in order to improve their accuracy. One important observation about the figures included in biomedical publications is that they are often composed of multiple subfigures or panels, each describing different methodologies or results. The use of these multimodal figures is a common practice in bioscience, as experimental results are graphically validated via multiple methodologies or procedures. Thus, for a better use of multimodal figures in document classification or retrieval tasks, as well as for providing the evidence source for derived assertions, it is important to automatically segment multimodal figures into subfigures and panels. This is a challenging task, however, as different panels can contain similar objects (i.e., barcharts and linecharts) with multiple layouts. Also, certain types of biomedical figures are text-heavy (e.g., DNA sequences and protein sequences images) and they differ from traditional images. As a result, classical image segmentation techniques based on low-level image features, such as edges or color, are not directly applicable to robustly partition multimodal figures into single modal panels. In this paper, we describe a robust solution for automatically identifying and segmenting unimodal panels from a multimodal figure. Our framework starts by robustly harvesting figure-caption pairs from biomedical articles. We base our approach on the observation that the document layout can be used to identify encoded figures and figure boundaries within PDF files. Taking into consideration the document layout allows us to correctly extract figures from the PDF document and associate their corresponding caption. We combine pixel-level representations of the extracted images with information gathered from their corresponding captions to estimate the number of panels in the figure. Thus, our approach simultaneously identifies the number of panels and the layout of figures. In order to evaluate the approach described here, we applied our system on documents containing protein-protein interactions (PPIs) and compared the results against a gold standard that was annotated by biologists. Experimental results showed that our automatic figure segmentation approach surpasses pure caption-based and image-based approaches, achieving a 96.64% accuracy. To allow for efficient retrieval of information, as well as to provide the basis for integration into document classification and retrieval systems among other, we further developed a web-based interface that lets users easily retrieve panels containing the terms specified in the user queries. PMID:24565394

Deep learning with non-medical training used for chest pathology identification

NASA Astrophysics Data System (ADS)

Bar, Yaniv; Diamant, Idit; Wolf, Lior; Greenspan, Hayit

2015-03-01

In this work, we examine the strength of deep learning approaches for pathology detection in chest radiograph data. Convolutional neural networks (CNN) deep architecture classification approaches have gained popularity due to their ability to learn mid and high level image representations. We explore the ability of a CNN to identify different types of pathologies in chest x-ray images. Moreover, since very large training sets are generally not available in the medical domain, we explore the feasibility of using a deep learning approach based on non-medical learning. We tested our algorithm on a dataset of 93 images. We use a CNN that was trained with ImageNet, a well-known large scale nonmedical image database. The best performance was achieved using a combination of features extracted from the CNN and a set of low-level features. We obtained an area under curve (AUC) of 0.93 for Right Pleural Effusion detection, 0.89 for Enlarged heart detection and 0.79 for classification between healthy and abnormal chest x-ray, where all pathologies are combined into one large class. This is a first-of-its-kind experiment that shows that deep learning with large scale non-medical image databases may be sufficient for general medical image recognition tasks.

Beam-hardening correction by a surface fitting and phase classification by a least square support vector machine approach for tomography images of geological samples

NASA Astrophysics Data System (ADS)

Khan, F.; Enzmann, F.; Kersten, M.

2015-12-01

In X-ray computed microtomography (μXCT) image processing is the most important operation prior to image analysis. Such processing mainly involves artefact reduction and image segmentation. We propose a new two-stage post-reconstruction procedure of an image of a geological rock core obtained by polychromatic cone-beam μXCT technology. In the first stage, the beam-hardening (BH) is removed applying a best-fit quadratic surface algorithm to a given image data set (reconstructed slice), which minimizes the BH offsets of the attenuation data points from that surface. The final BH-corrected image is extracted from the residual data, or the difference between the surface elevation values and the original grey-scale values. For the second stage, we propose using a least square support vector machine (a non-linear classifier algorithm) to segment the BH-corrected data as a pixel-based multi-classification task. A combination of the two approaches was used to classify a complex multi-mineral rock sample. The Matlab code for this approach is provided in the Appendix. A minor drawback is that the proposed segmentation algorithm may become computationally demanding in the case of a high dimensional training data set.

A semi-automatic method for quantification and classification of erythrocytes infected with malaria parasites in microscopic images.

PubMed

Díaz, Gloria; González, Fabio A; Romero, Eduardo

2009-04-01

Visual quantification of parasitemia in thin blood films is a very tedious, subjective and time-consuming task. This study presents an original method for quantification and classification of erythrocytes in stained thin blood films infected with Plasmodium falciparum. The proposed approach is composed of three main phases: a preprocessing step, which corrects luminance differences. A segmentation step that uses the normalized RGB color space for classifying pixels either as erythrocyte or background followed by an Inclusion-Tree representation that structures the pixel information into objects, from which erythrocytes are found. Finally, a two step classification process identifies infected erythrocytes and differentiates the infection stage, using a trained bank of classifiers. Additionally, user intervention is allowed when the approach cannot make a proper decision. Four hundred fifty malaria images were used for training and evaluating the method. Automatic identification of infected erythrocytes showed a specificity of 99.7% and a sensitivity of 94%. The infection stage was determined with an average sensitivity of 78.8% and average specificity of 91.2%.

Large Uptake of Titania and Iron Oxide Nanoparticles in the Nucleus of Lung Epithelial Cells as Measured by Raman Imaging and Multivariate Classification

PubMed Central

Ahlinder, Linnea; Ekstrand-Hammarström, Barbro; Geladi, Paul; Österlund, Lars

2013-01-01

It is a challenging task to characterize the biodistribution of nanoparticles in cells and tissue on a subcellular level. Conventional methods to study the interaction of nanoparticles with living cells rely on labeling techniques that either selectively stain the particles or selectively tag them with tracer molecules. In this work, Raman imaging, a label-free technique that requires no extensive sample preparation, was combined with multivariate classification to quantify the spatial distribution of oxide nanoparticles inside living lung epithelial cells (A549). Cells were exposed to TiO2 (titania) and/or α-FeO(OH) (goethite) nanoparticles at various incubation times (4 or 48 h). Using multivariate classification of hyperspectral Raman data with partial least-squares discriminant analysis, we show that a surprisingly large fraction of spectra, classified as belonging to the cell nucleus, show Raman bands associated with nanoparticles. Up to 40% of spectra from the cell nucleus show Raman bands associated with nanoparticles. Complementary transmission electron microscopy data for thin cell sections qualitatively support the conclusions. PMID:23870252

Direct Viewing of Dyslexics’ Compensatory Strategies in Speech in Noise Using Auditory Classification Images

PubMed Central

Varnet, Léo; Meunier, Fanny; Trollé, Gwendoline; Hoen, Michel

2016-01-01

A vast majority of dyslexic children exhibit a phonological deficit, particularly noticeable in phonemic identification or discrimination tasks. The gap in performance between dyslexic and normotypical listeners appears to decrease into adulthood, suggesting that some individuals with dyslexia develop compensatory strategies. Some dyslexic adults however remain impaired in more challenging listening situations such as in the presence of background noise. This paper addresses the question of the compensatory strategies employed, using the recently developed Auditory Classification Image (ACI) methodology. The results of 18 dyslexics taking part in a phoneme categorization task in noise were compared with those of 18 normotypical age-matched controls. By fitting a penalized Generalized Linear Model on the data of each participant, we obtained his/her ACI, a map of the time-frequency regions he/she relied on to perform the task. Even though dyslexics performed significantly less well than controls, we were unable to detect a robust difference between the mean ACIs of the two groups. This is partly due to the considerable heterogeneity in listening strategies among a subgroup of 7 low-performing dyslexics, as confirmed by a complementary analysis. When excluding these participants to restrict our comparison to the 11 dyslexics performing as well as their average-reading peers, we found a significant difference in the F3 onset of the first syllable, and a tendency of difference on the F4 onset, suggesting that these listeners can compensate for their deficit by relying upon additional allophonic cues. PMID:27100662

Direct Viewing of Dyslexics' Compensatory Strategies in Speech in Noise Using Auditory Classification Images.

PubMed

Varnet, Léo; Meunier, Fanny; Trollé, Gwendoline; Hoen, Michel

2016-01-01

A vast majority of dyslexic children exhibit a phonological deficit, particularly noticeable in phonemic identification or discrimination tasks. The gap in performance between dyslexic and normotypical listeners appears to decrease into adulthood, suggesting that some individuals with dyslexia develop compensatory strategies. Some dyslexic adults however remain impaired in more challenging listening situations such as in the presence of background noise. This paper addresses the question of the compensatory strategies employed, using the recently developed Auditory Classification Image (ACI) methodology. The results of 18 dyslexics taking part in a phoneme categorization task in noise were compared with those of 18 normotypical age-matched controls. By fitting a penalized Generalized Linear Model on the data of each participant, we obtained his/her ACI, a map of the time-frequency regions he/she relied on to perform the task. Even though dyslexics performed significantly less well than controls, we were unable to detect a robust difference between the mean ACIs of the two groups. This is partly due to the considerable heterogeneity in listening strategies among a subgroup of 7 low-performing dyslexics, as confirmed by a complementary analysis. When excluding these participants to restrict our comparison to the 11 dyslexics performing as well as their average-reading peers, we found a significant difference in the F3 onset of the first syllable, and a tendency of difference on the F4 onset, suggesting that these listeners can compensate for their deficit by relying upon additional allophonic cues.

Photoanthropometric face iridial proportions for age estimation: An investigation using features selected via a joint mutual information criterion.

PubMed

Borges, Díbio L; Vidal, Flávio B; Flores, Marta R P; Melani, Rodolfo F H; Guimarães, Marco A; Machado, Carlos E P

2018-03-01

Age assessment from images is of high interest in the forensic community because of the necessity to establish formal protocols to identify child pornography, child missing and abuses where visual evidences are the mostly admissible. Recently, photoanthropometric methods have been found useful for age estimation correlating facial proportions in image databases with samples of some age groups. Notwithstanding the advances, newer facial features and further analysis are needed to improve accuracy and establish larger applicability. In this investigation, frontal images of 1000 individuals (500 females, 500 males), equally distributed in five age groups (6, 10, 14, 18, 22 years old) were used in a 10 fold cross-validated experiment for three age thresholds classifications (<10, <14, <18 years old). A set of novel 40 features, based on a relation between landmark distances and the iris diameter, is proposed and joint mutual information is used to select the most relevant and complementary features for the classification task. In a civil image identification database with diverse ancestry, receiver operating characteristic (ROC) curves were plotted to verify accuracy, and the resultant AUCs achieved 0.971, 0.969, and 0.903 for the age classifications (<10, <14, <18 years old), respectively. These results add support to continuing research in age assessment from images using the metric approach. Still, larger samples are necessary to evaluate reliability in extensive conditions. Copyright © 2017 Elsevier B.V. All rights reserved.

Characterization of groups using composite kernels and multi-source fMRI analysis data: application to schizophrenia

PubMed Central

Castro, Eduardo; Martínez-Ramón, Manel; Pearlson, Godfrey; Sui, Jing; Calhoun, Vince D.

2011-01-01

Pattern classification of brain imaging data can enable the automatic detection of differences in cognitive processes of specific groups of interest. Furthermore, it can also give neuroanatomical information related to the regions of the brain that are most relevant to detect these differences by means of feature selection procedures, which are also well-suited to deal with the high dimensionality of brain imaging data. This work proposes the application of recursive feature elimination using a machine learning algorithm based on composite kernels to the classification of healthy controls and patients with schizophrenia. This framework, which evaluates nonlinear relationships between voxels, analyzes whole-brain fMRI data from an auditory task experiment that is segmented into anatomical regions and recursively eliminates the uninformative ones based on their relevance estimates, thus yielding the set of most discriminative brain areas for group classification. The collected data was processed using two analysis methods: the general linear model (GLM) and independent component analysis (ICA). GLM spatial maps as well as ICA temporal lobe and default mode component maps were then input to the classifier. A mean classification accuracy of up to 95% estimated with a leave-two-out cross-validation procedure was achieved by doing multi-source data classification. In addition, it is shown that the classification accuracy rate obtained by using multi-source data surpasses that reached by using single-source data, hence showing that this algorithm takes advantage of the complimentary nature of GLM and ICA. PMID:21723948

Applying machine learning classification techniques to automate sky object cataloguing

NASA Astrophysics Data System (ADS)

Fayyad, Usama M.; Doyle, Richard J.; Weir, W. Nick; Djorgovski, Stanislav

1993-08-01

We describe the application of an Artificial Intelligence machine learning techniques to the development of an automated tool for the reduction of a large scientific data set. The 2nd Mt. Palomar Northern Sky Survey is nearly completed. This survey provides comprehensive coverage of the northern celestial hemisphere in the form of photographic plates. The plates are being transformed into digitized images whose quality will probably not be surpassed in the next ten to twenty years. The images are expected to contain on the order of 107 galaxies and 108 stars. Astronomers wish to determine which of these sky objects belong to various classes of galaxies and stars. Unfortunately, the size of this data set precludes analysis in an exclusively manual fashion. Our approach is to develop a software system which integrates the functions of independently developed techniques for image processing and data classification. Digitized sky images are passed through image processing routines to identify sky objects and to extract a set of features for each object. These routines are used to help select a useful set of attributes for classifying sky objects. Then GID3 (Generalized ID3) and O-B Tree, two inductive learning techniques, learns classification decision trees from examples. These classifiers will then be applied to new data. These developmnent process is highly interactive, with astronomer input playing a vital role. Astronomers refine the feature set used to construct sky object descriptions, and evaluate the performance of the automated classification technique on new data. This paper gives an overview of the machine learning techniques with an emphasis on their general applicability, describes the details of our specific application, and reports the initial encouraging results. The results indicate that our machine learning approach is well-suited to the problem. The primary benefit of the approach is increased data reduction throughput. Another benefit is consistency of classification. The classification rules which are the product of the inductive learning techniques will form an objective, examinable basis for classifying sky objects. A final, not to be underestimated benefit is that astronomers will be freed from the tedium of an intensely visual task to pursue more challenging analysis and interpretation problems based on automatically catalogued data.

Object-Based Land Use Classification of Agricultural Land by Coupling Multi-Temporal Spectral Characteristics and Phenological Events in Germany

NASA Astrophysics Data System (ADS)

Knoefel, Patrick; Loew, Fabian; Conrad, Christopher

2015-04-01

Crop maps based on classification of remotely sensed data are of increased attendance in agricultural management. This induces a more detailed knowledge about the reliability of such spatial information. However, classification of agricultural land use is often limited by high spectral similarities of the studied crop types. More, spatially and temporally varying agro-ecological conditions can introduce confusion in crop mapping. Classification errors in crop maps in turn may have influence on model outputs, like agricultural production monitoring. One major goal of the PhenoS project ("Phenological structuring to determine optimal acquisition dates for Sentinel-2 data for field crop classification"), is the detection of optimal phenological time windows for land cover classification purposes. Since many crop species are spectrally highly similar, accurate classification requires the right selection of satellite images for a certain classification task. In the course of one growing season, phenological phases exist where crops are separable with higher accuracies. For this purpose, coupling of multi-temporal spectral characteristics and phenological events is promising. The focus of this study is set on the separation of spectrally similar cereal crops like winter wheat, barley, and rye of two test sites in Germany called "Harz/Central German Lowland" and "Demmin". However, this study uses object based random forest (RF) classification to investigate the impact of image acquisition frequency and timing on crop classification uncertainty by permuting all possible combinations of available RapidEye time series recorded on the test sites between 2010 and 2014. The permutations were applied to different segmentation parameters. Then, classification uncertainty was assessed and analysed, based on the probabilistic soft-output from the RF algorithm at the per-field basis. From this soft output, entropy was calculated as a spatial measure of classification uncertainty. The results indicate that uncertainty estimates provide a valuable addition to traditional accuracy assessments and helps the user to allocate error in crop maps.

Deep Multi-Task Learning for Tree Genera Classification

NASA Astrophysics Data System (ADS)

Ko, C.; Kang, J.; Sohn, G.

2018-05-01

The goal for our paper is to classify tree genera using airborne Light Detection and Ranging (LiDAR) data with Convolution Neural Network (CNN) - Multi-task Network (MTN) implementation. Unlike Single-task Network (STN) where only one task is assigned to the learning outcome, MTN is a deep learning architect for learning a main task (classification of tree genera) with other tasks (in our study, classification of coniferous and deciduous) simultaneously, with shared classification features. The main contribution of this paper is to improve classification accuracy from CNN-STN to CNN-MTN. This is achieved by introducing a concurrence loss (Lcd) to the designed MTN. This term regulates the overall network performance by minimizing the inconsistencies between the two tasks. Results show that we can increase the classification accuracy from 88.7 % to 91.0 % (from STN to MTN). The second goal of this paper is to solve the problem of small training sample size by multiple-view data generation. The motivation of this goal is to address one of the most common problems in implementing deep learning architecture, the insufficient number of training data. We address this problem by simulating training dataset with multiple-view approach. The promising results from this paper are providing a basis for classifying a larger number of dataset and number of classes in the future.

DecoFungi: a web application for automatic characterisation of dye decolorisation in fungal strains.

PubMed

Domínguez, César; Heras, Jónathan; Mata, Eloy; Pascual, Vico

2018-02-27

Fungi have diverse biotechnological applications in, among others, agriculture, bioenergy generation, or remediation of polluted soil and water. In this context, culture media based on color change in response to degradation of dyes are particularly relevant; but measuring dye decolorisation of fungal strains mainly relies on a visual and semiquantitative classification of color intensity changes. Such a classification is a subjective, time-consuming and difficult to reproduce process. DecoFungi is the first, at least up to the best of our knowledge, application to automatically characterise dye decolorisation level of fungal strains from images of inoculated plates. In order to deal with this task, DecoFungi employs a deep-learning model, accessible through a user-friendly web interface, with an accuracy of 96.5%. DecoFungi is an easy to use system for characterising dye decolorisation level of fungal strains from images of inoculated plates.

Automated Detection of Diabetic Retinopathy using Deep Learning.

PubMed

Lam, Carson; Yi, Darvin; Guo, Margaret; Lindsey, Tony

2018-01-01

Diabetic retinopathy is a leading cause of blindness among working-age adults. Early detection of this condition is critical for good prognosis. In this paper, we demonstrate the use of convolutional neural networks (CNNs) on color fundus images for the recognition task of diabetic retinopathy staging. Our network models achieved test metric performance comparable to baseline literature results, with validation sensitivity of 95%. We additionally explored multinomial classification models, and demonstrate that errors primarily occur in the misclassification of mild disease as normal due to the CNNs inability to detect subtle disease features. We discovered that preprocessing with contrast limited adaptive histogram equalization and ensuring dataset fidelity by expert verification of class labels improves recognition of subtle features. Transfer learning on pretrained GoogLeNet and AlexNet models from ImageNet improved peak test set accuracies to 74.5%, 68.8%, and 57.2% on 2-ary, 3-ary, and 4-ary classification models, respectively.

Local feature saliency classifier for real-time intrusion monitoring

NASA Astrophysics Data System (ADS)

Buch, Norbert; Velastin, Sergio A.

2014-07-01

We propose a texture saliency classifier to detect people in a video frame by identifying salient texture regions. The image is classified into foreground and background in real time. No temporal image information is used during the classification. The system is used for the task of detecting people entering a sterile zone, which is a common scenario for visual surveillance. Testing is performed on the Imagery Library for Intelligent Detection Systems sterile zone benchmark dataset of the United Kingdom's Home Office. The basic classifier is extended by fusing its output with simple motion information, which significantly outperforms standard motion tracking. A lower detection time can be achieved by combining texture classification with Kalman filtering. The fusion approach running at 10 fps gives the highest result of F1=0.92 for the 24-h test dataset. The paper concludes with a detailed analysis of the computation time required for the different parts of the algorithm.

Segmentation and analysis of mouse pituitary cells with graphic user interface (GUI)

NASA Astrophysics Data System (ADS)

González, Erika; Medina, Lucía.; Hautefeuille, Mathieu; Fiordelisio, Tatiana

2018-02-01

In this work we present a method to perform pituitary cell segmentation in image stacks acquired by fluorescence microscopy from pituitary slice preparations. Although there exist many procedures developed to achieve cell segmentation tasks, they are generally based on the edge detection and require high resolution images. However in the biological preparations that we worked on, the cells are not well defined as experts identify their intracellular calcium activity due to fluorescence intensity changes in different regions over time. This intensity changes were associated with time series over regions, and because they present a particular behavior they were used into a classification procedure in order to perform cell segmentation. Two logistic regression classifiers were implemented for the time series classification task using as features the area under the curve and skewness in the first classifier and skewness and kurtosis in the second classifier. Once we have found both decision boundaries in two different feature spaces by training using 120 time series, the decision boundaries were tested over 12 image stacks through a python graphical user interface (GUI), generating binary images where white pixels correspond to cells and the black ones to background. Results show that area-skewness classifier reduces the time an expert dedicates in locating cells by up to 75% in some stacks versus a 92% for the kurtosis-skewness classifier, this evaluated on the number of regions the method found. Due to the promising results, we expect that this method will be improved adding more relevant features to the classifier.

Characterization of healthy and osteoarthritic chondrocyte cell patterns on phase contrast CT images of the knee cartilage matrix

NASA Astrophysics Data System (ADS)

Nagarajan, Mahesh B.; Coan, Paola; Huber, Markus B.; Yang, Chien-Chun; Glaser, Christian; Reiser, Maximilian F.; Wismüller, Axel

2012-03-01

The current approach to evaluating cartilage degeneration at the knee joint requires visualization of the joint space on radiographic images where indirect cues such as joint space narrowing serve as markers for osteoarthritis. A recent novel approach to visualizing the knee cartilage matrix using phase contrast CT imaging (PCI-CT) was shown to allow direct examination of chondrocyte cell patterns and their subsequent correlation to osteoarthritis. This study aims to characterize chondrocyte cell patterns in the radial zone of the knee cartilage matrix in the presence and absence of osteoarthritic damage through both gray-level co-occurrence matrix (GLCM) derived texture features as well as Minkowski Functionals (MF). Thirteen GLCM and three MF texture features were extracted from 404 regions of interest (ROI) annotated on PCI images of healthy and osteoarthritic specimens of knee cartilage. These texture features were then used in a machine learning task to classify ROIs as healthy or osteoarthritic. A fuzzy k-nearest neighbor classifier was used and its performance was evaluated using the area under the ROC curve (AUC). The best classification performance was observed with the MF features 'perimeter' and 'Euler characteristic' and with GLCM correlation features (f3 and f13). With the experimental conditions used in this study, both Minkowski Functionals and GLCM achieved a high classification performance (AUC value of 0.97) in the task of distinguishing between health and osteoarthritic ROIs. These results show that such quantitative analysis of chondrocyte patterns in the knee cartilage matrix can distinguish between healthy and osteoarthritic tissue with high accuracy.

Pattern classification of kinematic and kinetic running data to distinguish gender, shod/barefoot and injury groups with feature ranking.

PubMed

Eskofier, Bjoern M; Kraus, Martin; Worobets, Jay T; Stefanyshyn, Darren J; Nigg, Benno M

2012-01-01

The identification of differences between groups is often important in biomechanics. This paper presents group classification tasks using kinetic and kinematic data from a prospective running injury study. Groups composed of gender, of shod/barefoot running and of runners who developed patellofemoral pain syndrome (PFPS) during the study, and asymptotic runners were classified. The features computed from the biomechanical data were deliberately chosen to be generic. Therefore, they were suited for different biomechanical measurements and classification tasks without adaptation to the input signals. Feature ranking was applied to reveal the relevance of each feature to the classification task. Data from 80 runners were analysed for gender and shod/barefoot classification, while 12 runners were investigated in the injury classification task. Gender groups could be differentiated with 84.7%, shod/barefoot running with 98.3%, and PFPS with 100% classification rate. For the latter group, one single variable could be identified that alone allowed discrimination.

Cell segmentation in phase contrast microscopy images via semi-supervised classification over optics-related features.

PubMed

Su, Hang; Yin, Zhaozheng; Huh, Seungil; Kanade, Takeo

2013-10-01

Phase-contrast microscopy is one of the most common and convenient imaging modalities to observe long-term multi-cellular processes, which generates images by the interference of lights passing through transparent specimens and background medium with different retarded phases. Despite many years of study, computer-aided phase contrast microscopy analysis on cell behavior is challenged by image qualities and artifacts caused by phase contrast optics. Addressing the unsolved challenges, the authors propose (1) a phase contrast microscopy image restoration method that produces phase retardation features, which are intrinsic features of phase contrast microscopy, and (2) a semi-supervised learning based algorithm for cell segmentation, which is a fundamental task for various cell behavior analysis. Specifically, the image formation process of phase contrast microscopy images is first computationally modeled with a dictionary of diffraction patterns; as a result, each pixel of a phase contrast microscopy image is represented by a linear combination of the bases, which we call phase retardation features. Images are then partitioned into phase-homogeneous atoms by clustering neighboring pixels with similar phase retardation features. Consequently, cell segmentation is performed via a semi-supervised classification technique over the phase-homogeneous atoms. Experiments demonstrate that the proposed approach produces quality segmentation of individual cells and outperforms previous approaches. Copyright © 2013 Elsevier B.V. All rights reserved.

Histopathological Breast Cancer Image Classification by Deep Neural Network Techniques Guided by Local Clustering.

PubMed

Nahid, Abdullah-Al; Mehrabi, Mohamad Ali; Kong, Yinan

2018-01-01

Breast Cancer is a serious threat and one of the largest causes of death of women throughout the world. The identification of cancer largely depends on digital biomedical photography analysis such as histopathological images by doctors and physicians. Analyzing histopathological images is a nontrivial task, and decisions from investigation of these kinds of images always require specialised knowledge. However, Computer Aided Diagnosis (CAD) techniques can help the doctor make more reliable decisions. The state-of-the-art Deep Neural Network (DNN) has been recently introduced for biomedical image analysis. Normally each image contains structural and statistical information. This paper classifies a set of biomedical breast cancer images (BreakHis dataset) using novel DNN techniques guided by structural and statistical information derived from the images. Specifically a Convolutional Neural Network (CNN), a Long-Short-Term-Memory (LSTM), and a combination of CNN and LSTM are proposed for breast cancer image classification. Softmax and Support Vector Machine (SVM) layers have been used for the decision-making stage after extracting features utilising the proposed novel DNN models. In this experiment the best Accuracy value of 91.00% is achieved on the 200x dataset, the best Precision value 96.00% is achieved on the 40x dataset, and the best F -Measure value is achieved on both the 40x and 100x datasets.

Hierarchy-associated semantic-rule inference framework for classifying indoor scenes

NASA Astrophysics Data System (ADS)

Yu, Dan; Liu, Peng; Ye, Zhipeng; Tang, Xianglong; Zhao, Wei

2016-03-01

Typically, the initial task of classifying indoor scenes is challenging, because the spatial layout and decoration of a scene can vary considerably. Recent efforts at classifying object relationships commonly depend on the results of scene annotation and predefined rules, making classification inflexible. Furthermore, annotation results are easily affected by external factors. Inspired by human cognition, a scene-classification framework was proposed using the empirically based annotation (EBA) and a match-over rule-based (MRB) inference system. The semantic hierarchy of images is exploited by EBA to construct rules empirically for MRB classification. The problem of scene classification is divided into low-level annotation and high-level inference from a macro perspective. Low-level annotation involves detecting the semantic hierarchy and annotating the scene with a deformable-parts model and a bag-of-visual-words model. In high-level inference, hierarchical rules are extracted to train the decision tree for classification. The categories of testing samples are generated from the parts to the whole. Compared with traditional classification strategies, the proposed semantic hierarchy and corresponding rules reduce the effect of a variable background and improve the classification performance. The proposed framework was evaluated on a popular indoor scene dataset, and the experimental results demonstrate its effectiveness.

Question analysis for Indonesian comparative question

NASA Astrophysics Data System (ADS)

Saelan, A.; Purwarianti, A.; Widyantoro, D. H.

2017-01-01

Information seeking is one of human needs today. Comparing things using search engine surely take more times than search only one thing. In this paper, we analyzed comparative questions for comparative question answering system. Comparative question is a question that comparing two or more entities. We grouped comparative questions into 5 types: selection between mentioned entities, selection between unmentioned entities, selection between any entity, comparison, and yes or no question. Then we extracted 4 types of information from comparative questions: entity, aspect, comparison, and constraint. We built classifiers for classification task and information extraction task. Features used for classification task are bag of words, whether for information extraction, we used lexical, 2 previous and following words lexical, and previous label as features. We tried 2 scenarios: classification first and extraction first. For classification first, we used classification result as a feature for extraction. Otherwise, for extraction first, we used extraction result as features for classification. We found that the result would be better if we do extraction first before classification. For the extraction task, classification using SMO gave the best result (88.78%), while for classification, it is better to use naïve bayes (82.35%).

Insights from Classifying Visual Concepts with Multiple Kernel Learning

PubMed Central

Binder, Alexander; Nakajima, Shinichi; Kloft, Marius; Müller, Christina; Samek, Wojciech; Brefeld, Ulf; Müller, Klaus-Robert; Kawanabe, Motoaki

2012-01-01

Combining information from various image features has become a standard technique in concept recognition tasks. However, the optimal way of fusing the resulting kernel functions is usually unknown in practical applications. Multiple kernel learning (MKL) techniques allow to determine an optimal linear combination of such similarity matrices. Classical approaches to MKL promote sparse mixtures. Unfortunately, 1-norm regularized MKL variants are often observed to be outperformed by an unweighted sum kernel. The main contributions of this paper are the following: we apply a recently developed non-sparse MKL variant to state-of-the-art concept recognition tasks from the application domain of computer vision. We provide insights on benefits and limits of non-sparse MKL and compare it against its direct competitors, the sum-kernel SVM and sparse MKL. We report empirical results for the PASCAL VOC 2009 Classification and ImageCLEF2010 Photo Annotation challenge data sets. Data sets (kernel matrices) as well as further information are available at http://doc.ml.tu-berlin.de/image_mkl/(Accessed 2012 Jun 25). PMID:22936970

CHIPS. Volume 27, Number 2, April-June 2009

DTIC Science & Technology

2009-07-01

unlimited 13. SUPPLEMENTARY NOTES The original document contains color images. 14. ABSTRACT 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17...Navy Network Enterprise 9 CARS Task Force Shortens Original Timeline – Reducing the number of Navy legacy networks to improve security and save...NETWARCOM and edited from the original article published in the Winter 2008-2009 edition of Info Domain. “A plan of action and milestones (POA&M) for

Applications of color machine vision in the agricultural and food industries

NASA Astrophysics Data System (ADS)

Zhang, Min; Ludas, Laszlo I.; Morgan, Mark T.; Krutz, Gary W.; Precetti, Cyrille J.

1999-01-01

Color is an important factor in Agricultural and the Food Industry. Agricultural or prepared food products are often grade by producers and consumers using color parameters. Color is used to estimate maturity, sort produce for defects, but also perform genetic screenings or make an aesthetic judgement. The task of sorting produce following a color scale is very complex, requires special illumination and training. Also, this task cannot be performed for long durations without fatigue and loss of accuracy. This paper describes a machine vision system designed to perform color classification in real-time. Applications for sorting a variety of agricultural products are included: e.g. seeds, meat, baked goods, plant and wood.FIrst the theory of color classification of agricultural and biological materials is introduced. Then, some tools for classifier development are presented. Finally, the implementation of the algorithm on real-time image processing hardware and example applications for industry is described. This paper also presented an image analysis algorithm and a prototype machine vision system which was developed for industry. This system will automatically locate the surface of some plants using digital camera and predict information such as size, potential value and type of this plant. The algorithm developed will be feasible for real-time identification in an industrial environment.

Differentiation of Glioblastoma and Lymphoma Using Feature Extraction and Support Vector Machine.

PubMed

Yang, Zhangjing; Feng, Piaopiao; Wen, Tian; Wan, Minghua; Hong, Xunning

2017-01-01

Differentiation of glioblastoma multiformes (GBMs) and lymphomas using multi-sequence magnetic resonance imaging (MRI) is an important task that is valuable for treatment planning. However, this task is a challenge because GBMs and lymphomas may have a similar appearance in MRI images. This similarity may lead to misclassification and could affect the treatment results. In this paper, we propose a semi-automatic method based on multi-sequence MRI to differentiate these two types of brain tumors. Our method consists of three steps: 1) the key slice is selected from 3D MRIs and region of interests (ROIs) are drawn around the tumor region; 2) different features are extracted based on prior clinical knowledge and validated using a t-test; and 3) features that are helpful for classification are used to build an original feature vector and a support vector machine is applied to perform classification. In total, 58 GBM cases and 37 lymphoma cases are used to validate our method. A leave-one-out crossvalidation strategy is adopted in our experiments. The global accuracy of our method was determined as 96.84%, which indicates that our method is effective for the differentiation of GBM and lymphoma and can be applied in clinical diagnosis. Copyright© Bentham Science Publishers; For any queries, please email at epub@benthamscience.org.

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.

Multi-scale Gaussian representation and outline-learning based cell image segmentation.

PubMed

Farhan, Muhammad; Ruusuvuori, Pekka; Emmenlauer, Mario; Rämö, Pauli; Dehio, Christoph; Yli-Harja, Olli

2013-01-01

High-throughput genome-wide screening to study gene-specific functions, e.g. for drug discovery, demands fast automated image analysis methods to assist in unraveling the full potential of such studies. Image segmentation is typically at the forefront of such analysis as the performance of the subsequent steps, for example, cell classification, cell tracking etc., often relies on the results of segmentation. We present a cell cytoplasm segmentation framework which first separates cell cytoplasm from image background using novel approach of image enhancement and coefficient of variation of multi-scale Gaussian scale-space representation. A novel outline-learning based classification method is developed using regularized logistic regression with embedded feature selection which classifies image pixels as outline/non-outline to give cytoplasm outlines. Refinement of the detected outlines to separate cells from each other is performed in a post-processing step where the nuclei segmentation is used as contextual information. We evaluate the proposed segmentation methodology using two challenging test cases, presenting images with completely different characteristics, with cells of varying size, shape, texture and degrees of overlap. The feature selection and classification framework for outline detection produces very simple sparse models which use only a small subset of the large, generic feature set, that is, only 7 and 5 features for the two cases. Quantitative comparison of the results for the two test cases against state-of-the-art methods show that our methodology outperforms them with an increase of 4-9% in segmentation accuracy with maximum accuracy of 93%. Finally, the results obtained for diverse datasets demonstrate that our framework not only produces accurate segmentation but also generalizes well to different segmentation tasks.

Multi-scale Gaussian representation and outline-learning based cell image segmentation

PubMed Central

2013-01-01

Background High-throughput genome-wide screening to study gene-specific functions, e.g. for drug discovery, demands fast automated image analysis methods to assist in unraveling the full potential of such studies. Image segmentation is typically at the forefront of such analysis as the performance of the subsequent steps, for example, cell classification, cell tracking etc., often relies on the results of segmentation. Methods We present a cell cytoplasm segmentation framework which first separates cell cytoplasm from image background using novel approach of image enhancement and coefficient of variation of multi-scale Gaussian scale-space representation. A novel outline-learning based classification method is developed using regularized logistic regression with embedded feature selection which classifies image pixels as outline/non-outline to give cytoplasm outlines. Refinement of the detected outlines to separate cells from each other is performed in a post-processing step where the nuclei segmentation is used as contextual information. Results and conclusions We evaluate the proposed segmentation methodology using two challenging test cases, presenting images with completely different characteristics, with cells of varying size, shape, texture and degrees of overlap. The feature selection and classification framework for outline detection produces very simple sparse models which use only a small subset of the large, generic feature set, that is, only 7 and 5 features for the two cases. Quantitative comparison of the results for the two test cases against state-of-the-art methods show that our methodology outperforms them with an increase of 4-9% in segmentation accuracy with maximum accuracy of 93%. Finally, the results obtained for diverse datasets demonstrate that our framework not only produces accurate segmentation but also generalizes well to different segmentation tasks. PMID:24267488

Free classification of regional dialects of American English.

PubMed

Clopper, Cynthia G; Pisoni, David B

2007-07-01

Recent studies have found that naïve listeners perform poorly in forced-choice dialect categorization tasks. However, the listeners' error patterns in these tasks reveal systematic confusions between phonologically similar dialects. In the present study, a free classification procedure was used to measure the perceptual similarity structure of regional dialect variation in the United States. In two experiments, participants listened to a set of short English sentences produced by male talkers only (Experiment 1) and by male and female talkers (Experiment 2). The listeners were instructed to group the talkers by regional dialect into as many groups as they wanted with as many talkers in each group as they wished. Multidimensional scaling analyses of the data revealed three primary dimensions of perceptual similarity (linguistic markedness, geography, and gender). In addition, a comparison of the results obtained from the free classification task to previous results using the same stimulus materials in six-alternative forced-choice categorization tasks revealed that response biases in the six-alternative task were reduced or eliminated in the free classification task. Thus, the results obtained with the free classification task in the current study provided further evidence that the underlying structure of perceptual dialect category representations reflects important linguistic and sociolinguistic factors.

Semantic segmentation of forest stands of pure species combining airborne lidar data and very high resolution multispectral imagery

NASA Astrophysics Data System (ADS)

Dechesne, Clément; Mallet, Clément; Le Bris, Arnaud; Gouet-Brunet, Valérie

2017-04-01

Forest stands are the basic units for forest inventory and mapping. Stands are defined as large forested areas (e.g., ⩾ 2 ha) of homogeneous tree species composition and age. Their accurate delineation is usually performed by human operators through visual analysis of very high resolution (VHR) infra-red images. This task is tedious, highly time consuming, and should be automated for scalability and efficient updating purposes. In this paper, a method based on the fusion of airborne lidar data and VHR multispectral images is proposed for the automatic delineation of forest stands containing one dominant species (purity superior to 75%). This is the key preliminary task for forest land-cover database update. The multispectral images give information about the tree species whereas 3D lidar point clouds provide geometric information on the trees and allow their individual extraction. Multi-modal features are computed, both at pixel and object levels: the objects are individual trees extracted from lidar data. A supervised classification is then performed at the object level in order to coarsely discriminate the existing tree species in each area of interest. The classification results are further processed to obtain homogeneous areas with smooth borders by employing an energy minimum framework, where additional constraints are joined to form the energy function. The experimental results show that the proposed method provides very satisfactory results both in terms of stand labeling and delineation (overall accuracy ranges between 84 % and 99 %).

NutriNet: A Deep Learning Food and Drink Image Recognition System for Dietary Assessment.

PubMed

Mezgec, Simon; Koroušić Seljak, Barbara

2017-06-27

Automatic food image recognition systems are alleviating the process of food-intake estimation and dietary assessment. However, due to the nature of food images, their recognition is a particularly challenging task, which is why traditional approaches in the field have achieved a low classification accuracy. Deep neural networks have outperformed such solutions, and we present a novel approach to the problem of food and drink image detection and recognition that uses a newly-defined deep convolutional neural network architecture, called NutriNet. This architecture was tuned on a recognition dataset containing 225,953 512 × 512 pixel images of 520 different food and drink items from a broad spectrum of food groups, on which we achieved a classification accuracy of 86 . 72 % , along with an accuracy of 94 . 47 % on a detection dataset containing 130 , 517 images. We also performed a real-world test on a dataset of self-acquired images, combined with images from Parkinson's disease patients, all taken using a smartphone camera, achieving a top-five accuracy of 55 % , which is an encouraging result for real-world images. Additionally, we tested NutriNet on the University of Milano-Bicocca 2016 (UNIMIB2016) food image dataset, on which we improved upon the provided baseline recognition result. An online training component was implemented to continually fine-tune the food and drink recognition model on new images. The model is being used in practice as part of a mobile app for the dietary assessment of Parkinson's disease patients.

Risk assessments using the Strain Index and the TLV for HAL, Part I: Task and multi-task job exposure classifications.

PubMed

Kapellusch, Jay M; Bao, Stephen S; Silverstein, Barbara A; Merryweather, Andrew S; Thiese, Mathew S; Hegmann, Kurt T; Garg, Arun

2017-12-01

The Strain Index (SI) and the American Conference of Governmental Industrial Hygienists (ACGIH) Threshold Limit Value for Hand Activity Level (TLV for HAL) use different constituent variables to quantify task physical exposures. Similarly, time-weighted-average (TWA), Peak, and Typical exposure techniques to quantify physical exposure from multi-task jobs make different assumptions about each task's contribution to the whole job exposure. Thus, task and job physical exposure classifications differ depending upon which model and technique are used for quantification. This study examines exposure classification agreement, disagreement, correlation, and magnitude of classification differences between these models and techniques. Data from 710 multi-task job workers performing 3,647 tasks were analyzed using the SI and TLV for HAL models, as well as with the TWA, Typical and Peak job exposure techniques. Physical exposures were classified as low, medium, and high using each model's recommended, or a priori limits. Exposure classification agreement and disagreement between models (SI, TLV for HAL) and between job exposure techniques (TWA, Typical, Peak) were described and analyzed. Regardless of technique, the SI classified more tasks as high exposure than the TLV for HAL, and the TLV for HAL classified more tasks as low exposure. The models agreed on 48.5% of task classifications (kappa = 0.28) with 15.5% of disagreement between low and high exposure categories. Between-technique (i.e., TWA, Typical, Peak) agreement ranged from 61-93% (kappa: 0.16-0.92) depending on whether the SI or TLV for HAL was used. There was disagreement between the SI and TLV for HAL and between the TWA, Typical and Peak techniques. Disagreement creates uncertainty for job design, job analysis, risk assessments, and developing interventions. Task exposure classifications from the SI and TLV for HAL might complement each other. However, TWA, Typical, and Peak job exposure techniques all have limitations. Part II of this article examines whether the observed differences between these models and techniques produce different exposure-response relationships for predicting prevalence of carpal tunnel syndrome.

AutoBD: Automated Bi-Level Description for Scalable Fine-Grained Visual Categorization.

PubMed

Yao, Hantao; Zhang, Shiliang; Yan, Chenggang; Zhang, Yongdong; Li, Jintao; Tian, Qi

Compared with traditional image classification, fine-grained visual categorization is a more challenging task, because it targets to classify objects belonging to the same species, e.g. , classify hundreds of birds or cars. In the past several years, researchers have made many achievements on this topic. However, most of them are heavily dependent on the artificial annotations, e.g., bounding boxes, part annotations, and so on . The requirement of artificial annotations largely hinders the scalability and application. Motivated to release such dependence, this paper proposes a robust and discriminative visual description named Automated Bi-level Description (AutoBD). "Bi-level" denotes two complementary part-level and object-level visual descriptions, respectively. AutoBD is "automated," because it only requires the image-level labels of training images and does not need any annotations for testing images. Compared with the part annotations labeled by the human, the image-level labels can be easily acquired, which thus makes AutoBD suitable for large-scale visual categorization. Specifically, the part-level description is extracted by identifying the local region saliently representing the visual distinctiveness. The object-level description is extracted from object bounding boxes generated with a co-localization algorithm. Although only using the image-level labels, AutoBD outperforms the recent studies on two public benchmark, i.e. , classification accuracy achieves 81.6% on CUB-200-2011 and 88.9% on Car-196, respectively. On the large-scale Birdsnap data set, AutoBD achieves the accuracy of 68%, which is currently the best performance to the best of our knowledge.Compared with traditional image classification, fine-grained visual categorization is a more challenging task, because it targets to classify objects belonging to the same species, e.g. , classify hundreds of birds or cars. In the past several years, researchers have made many achievements on this topic. However, most of them are heavily dependent on the artificial annotations, e.g., bounding boxes, part annotations, and so on . The requirement of artificial annotations largely hinders the scalability and application. Motivated to release such dependence, this paper proposes a robust and discriminative visual description named Automated Bi-level Description (AutoBD). "Bi-level" denotes two complementary part-level and object-level visual descriptions, respectively. AutoBD is "automated," because it only requires the image-level labels of training images and does not need any annotations for testing images. Compared with the part annotations labeled by the human, the image-level labels can be easily acquired, which thus makes AutoBD suitable for large-scale visual categorization. Specifically, the part-level description is extracted by identifying the local region saliently representing the visual distinctiveness. The object-level description is extracted from object bounding boxes generated with a co-localization algorithm. Although only using the image-level labels, AutoBD outperforms the recent studies on two public benchmark, i.e. , classification accuracy achieves 81.6% on CUB-200-2011 and 88.9% on Car-196, respectively. On the large-scale Birdsnap data set, AutoBD achieves the accuracy of 68%, which is currently the best performance to the best of our knowledge.

Structural brain changes versus self-report: machine-learning classification of chronic fatigue syndrome patients.

PubMed

Sevel, Landrew S; Boissoneault, Jeff; Letzen, Janelle E; Robinson, Michael E; Staud, Roland

2018-05-30

Chronic fatigue syndrome (CFS) is a disorder associated with fatigue, pain, and structural/functional abnormalities seen during magnetic resonance brain imaging (MRI). Therefore, we evaluated the performance of structural MRI (sMRI) abnormalities in the classification of CFS patients versus healthy controls and compared it to machine learning (ML) classification based upon self-report (SR). Participants included 18 CFS patients and 15 healthy controls (HC). All subjects underwent T1-weighted sMRI and provided visual analogue-scale ratings of fatigue, pain intensity, anxiety, depression, anger, and sleep quality. sMRI data were segmented using FreeSurfer and 61 regions based on functional and structural abnormalities previously reported in patients with CFS. Classification was performed in RapidMiner using a linear support vector machine and bootstrap optimism correction. We compared ML classifiers based on (1) 61 a priori sMRI regional estimates and (2) SR ratings. The sMRI model achieved 79.58% classification accuracy. The SR (accuracy = 95.95%) outperformed both sMRI models. Estimates from multiple brain areas related to cognition, emotion, and memory contributed strongly to group classification. This is the first ML-based group classification of CFS. Our findings suggest that sMRI abnormalities are useful for discriminating CFS patients from HC, but SR ratings remain most effective in classification tasks.

A new classification scheme of plastic wastes based upon recycling labels.

PubMed

Özkan, Kemal; Ergin, Semih; Işık, Şahin; Işıklı, Idil

2015-01-01

Since recycling of materials is widely assumed to be environmentally and economically beneficial, reliable sorting and processing of waste packaging materials such as plastics is very important for recycling with high efficiency. An automated system that can quickly categorize these materials is certainly needed for obtaining maximum classification while maintaining high throughput. In this paper, first of all, the photographs of the plastic bottles have been taken and several preprocessing steps were carried out. The first preprocessing step is to extract the plastic area of a bottle from the background. Then, the morphological image operations are implemented. These operations are edge detection, noise removal, hole removing, image enhancement, and image segmentation. These morphological operations can be generally defined in terms of the combinations of erosion and dilation. The effect of bottle color as well as label are eliminated using these operations. Secondly, the pixel-wise intensity values of the plastic bottle images have been used together with the most popular subspace and statistical feature extraction methods to construct the feature vectors in this study. Only three types of plastics are considered due to higher existence ratio of them than the other plastic types in the world. The decision mechanism consists of five different feature extraction methods including as Principal Component Analysis (PCA), Kernel PCA (KPCA), Fisher's Linear Discriminant Analysis (FLDA), Singular Value Decomposition (SVD) and Laplacian Eigenmaps (LEMAP) and uses a simple experimental setup with a camera and homogenous backlighting. Due to the giving global solution for a classification problem, Support Vector Machine (SVM) is selected to achieve the classification task and majority voting technique is used as the decision mechanism. This technique equally weights each classification result and assigns the given plastic object to the class that the most classification results agree on. The proposed classification scheme provides high accuracy rate, and also it is able to run in real-time applications. It can automatically classify the plastic bottle types with approximately 90% recognition accuracy. Besides this, the proposed methodology yields approximately 96% classification rate for the separation of PET or non-PET plastic types. It also gives 92% accuracy for the categorization of non-PET plastic types into HPDE or PP. Copyright © 2014 Elsevier Ltd. All rights reserved.

New approach for cognitive analysis and understanding of medical patterns and visualizations

NASA Astrophysics Data System (ADS)

Ogiela, Marek R.; Tadeusiewicz, Ryszard

2003-11-01

This paper presents new opportunities for applying linguistic description of the picture merit content and AI methods to undertake tasks of the automatic understanding of images semantics in intelligent medical information systems. A successful obtaining of the crucial semantic content of the medical image may contribute considerably to the creation of new intelligent multimedia cognitive medical systems. Thanks to the new idea of cognitive resonance between stream of the data extracted from the image using linguistic methods and expectations taken from the representaion of the medical knowledge, it is possible to understand the merit content of the image even if teh form of the image is very different from any known pattern. This article proves that structural techniques of artificial intelligence may be applied in the case of tasks related to automatic classification and machine perception based on semantic pattern content in order to determine the semantic meaning of the patterns. In the paper are described some examples presenting ways of applying such techniques in the creation of cognitive vision systems for selected classes of medical images. On the base of scientific research described in the paper we try to build some new systems for collecting, storing, retrieving and intelligent interpreting selected medical images especially obtained in radiological and MRI examinations.

Classification of teeth in cone-beam CT using deep convolutional neural network.

PubMed

Miki, Yuma; Muramatsu, Chisako; Hayashi, Tatsuro; Zhou, Xiangrong; Hara, Takeshi; Katsumata, Akitoshi; Fujita, Hiroshi

2017-01-01

Dental records play an important role in forensic identification. To this end, postmortem dental findings and teeth conditions are recorded in a dental chart and compared with those of antemortem records. However, most dentists are inexperienced at recording the dental chart for corpses, and it is a physically and mentally laborious task, especially in large scale disasters. Our goal is to automate the dental filing process by using dental x-ray images. In this study, we investigated the application of a deep convolutional neural network (DCNN) for classifying tooth types on dental cone-beam computed tomography (CT) images. Regions of interest (ROIs) including single teeth were extracted from CT slices. Fifty two CT volumes were randomly divided into 42 training and 10 test cases, and the ROIs obtained from the training cases were used for training the DCNN. For examining the sampling effect, random sampling was performed 3 times, and training and testing were repeated. We used the AlexNet network architecture provided in the Caffe framework, which consists of 5 convolution layers, 3 pooling layers, and 2 full connection layers. For reducing the overtraining effect, we augmented the data by image rotation and intensity transformation. The test ROIs were classified into 7 tooth types by the trained network. The average classification accuracy using the augmented training data by image rotation and intensity transformation was 88.8%. Compared with the result without data augmentation, data augmentation resulted in an approximately 5% improvement in classification accuracy. This indicates that the further improvement can be expected by expanding the CT dataset. Unlike the conventional methods, the proposed method is advantageous in obtaining high classification accuracy without the need for precise tooth segmentation. The proposed tooth classification method can be useful in automatic filing of dental charts for forensic identification. Copyright © 2016 Elsevier Ltd. All rights reserved.

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.

Cytopathological image analysis using deep-learning networks in microfluidic microscopy.

PubMed

Gopakumar, G; Hari Babu, K; Mishra, Deepak; Gorthi, Sai Siva; Sai Subrahmanyam, Gorthi R K

2017-01-01

Cytopathologic testing is one of the most critical steps in the diagnosis of diseases, including cancer. However, the task is laborious and demands skill. Associated high cost and low throughput drew considerable interest in automating the testing process. Several neural network architectures were designed to provide human expertise to machines. In this paper, we explore and propose the feasibility of using deep-learning networks for cytopathologic analysis by performing the classification of three important unlabeled, unstained leukemia cell lines (K562, MOLT, and HL60). The cell images used in the classification are captured using a low-cost, high-throughput cell imaging technique: microfluidics-based imaging flow cytometry. We demonstrate that without any conventional fine segmentation followed by explicit feature extraction, the proposed deep-learning algorithms effectively classify the coarsely localized cell lines. We show that the designed deep belief network as well as the deeply pretrained convolutional neural network outperform the conventionally used decision systems and are important in the medical domain, where the availability of labeled data is limited for training. We hope that our work enables the development of a clinically significant high-throughput microfluidic microscopy-based tool for disease screening/triaging, especially in resource-limited settings.

Landsat real-time processing

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

Davis, E.L.

A novel method for performing real-time acquisition and processing Landsat/EROS data covers all aspects including radiometric and geometric corrections of multispectral scanner or return-beam vidicon inputs, image enhancement, statistical analysis, feature extraction, and classification. Radiometric transformations include bias/gain adjustment, noise suppression, calibration, scan angle compensation, and illumination compensation, including topography and atmospheric effects. Correction or compensation for geometric distortion includes sensor-related distortions, such as centering, skew, size, scan nonlinearity, radial symmetry, and tangential symmetry. Also included are object image-related distortions such as aspect angle (altitude), scale distortion (altitude), terrain relief, and earth curvature. Ephemeral corrections are also applied to compensatemore » for satellite forward movement, earth rotation, altitude variations, satellite vibration, and mirror scan velocity. Image enhancement includes high-pass, low-pass, and Laplacian mask filtering and data restoration for intermittent losses. Resource classification is provided by statistical analysis including histograms, correlational analysis, matrix manipulations, and determination of spectral responses. Feature extraction includes spatial frequency analysis, which is used in parallel discriminant functions in each array processor for rapid determination. The technique uses integrated parallel array processors that decimate the tasks concurrently under supervision of a control processor. The operator-machine interface is optimized for programming ease and graphics image windowing.« less

Towards improved characterization of northern wetlands (or other landscapes) by remote sensing - a rapid approach to collect ground truth data

NASA Astrophysics Data System (ADS)

Gålfalk, Magnus; Karlson, Martin; Crill, Patrick; Bastviken, David

2017-04-01

The calibration and validation of remote sensing land cover products is highly dependent on accurate ground truth data, which are costly and practically challenging to collect. This study evaluates a novel and efficient alternative to field surveys and UAV imaging commonly applied for this task. The method consists of i) a light weight, water proof, remote controlled RGB-camera mounted on an extendable monopod used for acquiring wide-field images of the ground from a height of 4.5 meters, and ii) a script for semi-automatic image classification. In the post-processing, the wide-field images are corrected for optical distortion and geometrically rectified so that the spatial resolution is the same over the surface area used for classification. The script distinguishes land surface components by color, brightness and spatial variability. The method was evaluated in wetland areas located around Abisko, northern Sweden. Proportional estimates of the six main surface components in the wetlands (wet and dry Sphagnum, shrub, grass, water, rock) were derived for 200 images, equivalent to 10 × 10 m field plots. These photo plots were then used as calibration data for a regional scale satellite based classification which separates the six wetland surface components using a Sentinel-1 time series. The method presented in this study is accurate, rapid, robust and cost efficient in comparison to field surveys (time consuming) and drone mapping (which require low wind speeds and no rain, suffer from battery limited flight times, have potential GPS/compass errors far north, and in some areas are prohibited by law).

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.

Classification of bifurcations regions in IVOCT images using support vector machine and artificial neural network models

NASA Astrophysics Data System (ADS)

Porto, C. D. N.; Costa Filho, C. F. F.; Macedo, M. M. G.; Gutierrez, M. A.; Costa, M. G. F.

2017-03-01

Studies in intravascular optical coherence tomography (IV-OCT) have demonstrated the importance of coronary bifurcation regions in intravascular medical imaging analysis, as plaques are more likely to accumulate in this region leading to coronary disease. A typical IV-OCT pullback acquires hundreds of frames, thus developing an automated tool to classify the OCT frames as bifurcation or non-bifurcation can be an important step to speed up OCT pullbacks analysis and assist automated methods for atherosclerotic plaque quantification. In this work, we evaluate the performance of two state-of-the-art classifiers, SVM and Neural Networks in the bifurcation classification task. The study included IV-OCT frames from 9 patients. In order to improve classification performance, we trained and tested the SVM with different parameters by means of a grid search and different stop criteria were applied to the Neural Network classifier: mean square error, early stop and regularization. Different sets of features were tested, using feature selection techniques: PCA, LDA and scalar feature selection with correlation. Training and test were performed in sets with a maximum of 1460 OCT frames. We quantified our results in terms of false positive rate, true positive rate, accuracy, specificity, precision, false alarm, f-measure and area under ROC curve. Neural networks obtained the best classification accuracy, 98.83%, overcoming the results found in literature. Our methods appear to offer a robust and reliable automated classification of OCT frames that might assist physicians indicating potential frames to analyze. Methods for improving neural networks generalization have increased the classification performance.

A distributed pipeline for DIDSON data processing

USGS Publications Warehouse

Li, Liling; Danner, Tyler; Eickholt, Jesse; McCann, Erin L.; Pangle, Kevin; Johnson, Nicholas

2018-01-01

Technological advances in the field of ecology allow data on ecological systems to be collected at high resolution, both temporally and spatially. Devices such as Dual-frequency Identification Sonar (DIDSON) can be deployed in aquatic environments for extended periods and easily generate several terabytes of underwater surveillance data which may need to be processed multiple times. Due to the large amount of data generated and need for flexibility in processing, a distributed pipeline was constructed for DIDSON data making use of the Hadoop ecosystem. The pipeline is capable of ingesting raw DIDSON data, transforming the acoustic data to images, filtering the images, detecting and extracting motion, and generating feature data for machine learning and classification. All of the tasks in the pipeline can be run in parallel and the framework allows for custom processing. Applications of the pipeline include monitoring migration times, determining the presence of a particular species, estimating population size and other fishery management tasks.

[Medical computer-aided detection method based on deep learning].

PubMed

Tao, Pan; Fu, Zhongliang; Zhu, Kai; Wang, Lili

2018-03-01

This paper performs a comprehensive study on the computer-aided detection for the medical diagnosis with deep learning. Based on the region convolution neural network and the prior knowledge of target, this algorithm uses the region proposal network, the region of interest pooling strategy, introduces the multi-task loss function: classification loss, bounding box localization loss and object rotation loss, and optimizes it by end-to-end. For medical image it locates the target automatically, and provides the localization result for the next stage task of segmentation. For the detection of left ventricular in echocardiography, proposed additional landmarks such as mitral annulus, endocardial pad and apical position, were used to estimate the left ventricular posture effectively. In order to verify the robustness and effectiveness of the algorithm, the experimental data of ultrasonic and nuclear magnetic resonance images are selected. Experimental results show that the algorithm is fast, accurate and effective.

Deep learning for neuroimaging: a validation study.

PubMed

Plis, Sergey M; Hjelm, Devon R; Salakhutdinov, Ruslan; Allen, Elena A; Bockholt, Henry J; Long, Jeffrey D; Johnson, Hans J; Paulsen, Jane S; Turner, Jessica A; Calhoun, Vince D

2014-01-01

Deep learning methods have recently made notable advances in the tasks of classification and representation learning. These tasks are important for brain imaging and neuroscience discovery, making the methods attractive for porting to a neuroimager's toolbox. Success of these methods is, in part, explained by the flexibility of deep learning models. However, this flexibility makes the process of porting to new areas a difficult parameter optimization problem. In this work we demonstrate our results (and feasible parameter ranges) in application of deep learning methods to structural and functional brain imaging data. These methods include deep belief networks and their building block the restricted Boltzmann machine. We also describe a novel constraint-based approach to visualizing high dimensional data. We use it to analyze the effect of parameter choices on data transformations. Our results show that deep learning methods are able to learn physiologically important representations and detect latent relations in neuroimaging data.

Pure word deafness with auditory object agnosia after bilateral lesion of the superior temporal sulcus.

PubMed

Gutschalk, Alexander; Uppenkamp, Stefan; Riedel, Bernhard; Bartsch, Andreas; Brandt, Tobias; Vogt-Schaden, Marlies

2015-12-01

Based on results from functional imaging, cortex along the superior temporal sulcus (STS) has been suggested to subserve phoneme and pre-lexical speech perception. For vowel classification, both superior temporal plane (STP) and STS areas have been suggested relevant. Lesion of bilateral STS may conversely be expected to cause pure word deafness and possibly also impaired vowel classification. Here we studied a patient with bilateral STS lesions caused by ischemic strokes and relatively intact medial STPs to characterize the behavioral consequences of STS loss. The patient showed severe deficits in auditory speech perception, whereas his speech production was fluent and communication by written speech was grossly intact. Auditory-evoked fields in the STP were within normal limits on both sides, suggesting that major parts of the auditory cortex were functionally intact. Further studies showed that the patient had normal hearing thresholds and only mild disability in tests for telencephalic hearing disorder. Prominent deficits were discovered in an auditory-object classification task, where the patient performed four standard deviations below the control group. In marked contrast, performance in a vowel-classification task was intact. Auditory evoked fields showed enhanced responses for vowels compared to matched non-vowels within normal limits. Our results are consistent with the notion that cortex along STS is important for auditory speech perception, although it does not appear to be entirely speech specific. Formant analysis and single vowel classification, however, appear to be already implemented in auditory cortex on the STP. Copyright © 2015 Elsevier Ltd. All rights reserved.

On-board multispectral classification study. Volume 2: Supplementary tasks. [adaptive control

NASA Technical Reports Server (NTRS)

Ewalt, D.

1979-01-01

The operational tasks of the onboard multispectral classification study were defined. These tasks include: sensing characteristics for future space applications; information adaptive systems architectural approaches; data set selection criteria; and onboard functional requirements for interfacing with global positioning satellites.

HCP: A Flexible CNN Framework for Multi-label Image Classification.

PubMed

Wei, Yunchao; Xia, Wei; Lin, Min; Huang, Junshi; Ni, Bingbing; Dong, Jian; Zhao, Yao; Yan, Shuicheng

2015-10-26

Convolutional Neural Network (CNN) has demonstrated promising performance in single-label image classification tasks. However, how CNN best copes with multi-label images still remains an open problem, mainly due to the complex underlying object layouts and insufficient multi-label training images. In this work, we propose a flexible deep CNN infrastructure, called Hypotheses-CNN-Pooling (HCP), where an arbitrary number of object segment hypotheses are taken as the inputs, then a shared CNN is connected with each hypothesis, and finally the CNN output results from different hypotheses are aggregated with max pooling to produce the ultimate multi-label predictions. Some unique characteristics of this flexible deep CNN infrastructure include: 1) no ground-truth bounding box information is required for training; 2) the whole HCP infrastructure is robust to possibly noisy and/or redundant hypotheses; 3) the shared CNN is flexible and can be well pre-trained with a large-scale single-label image dataset, e.g., ImageNet; and 4) it may naturally output multi-label prediction results. Experimental results on Pascal VOC 2007 and VOC 2012 multi-label image datasets well demonstrate the superiority of the proposed HCP infrastructure over other state-of-the-arts. In particular, the mAP reaches 90.5% by HCP only and 93.2% after the fusion with our complementary result in [44] based on hand-crafted features on the VOC 2012 dataset.

Object Detection and Classification by Decision-Level Fusion for Intelligent Vehicle Systems.

PubMed

Oh, Sang-Il; Kang, Hang-Bong

2017-01-22

To understand driving environments effectively, it is important to achieve accurate detection and classification of objects detected by sensor-based intelligent vehicle systems, which are significantly important tasks. Object detection is performed for the localization of objects, whereas object classification recognizes object classes from detected object regions. For accurate object detection and classification, fusing multiple sensor information into a key component of the representation and perception processes is necessary. In this paper, we propose a new object-detection and classification method using decision-level fusion. We fuse the classification outputs from independent unary classifiers, such as 3D point clouds and image data using a convolutional neural network (CNN). The unary classifiers for the two sensors are the CNN with five layers, which use more than two pre-trained convolutional layers to consider local to global features as data representation. To represent data using convolutional layers, we apply region of interest (ROI) pooling to the outputs of each layer on the object candidate regions generated using object proposal generation to realize color flattening and semantic grouping for charge-coupled device and Light Detection And Ranging (LiDAR) sensors. We evaluate our proposed method on a KITTI benchmark dataset to detect and classify three object classes: cars, pedestrians and cyclists. The evaluation results show that the proposed method achieves better performance than the previous methods. Our proposed method extracted approximately 500 proposals on a 1226 × 370 image, whereas the original selective search method extracted approximately 10 6 × n proposals. We obtained classification performance with 77.72% mean average precision over the entirety of the classes in the moderate detection level of the KITTI benchmark dataset.

Object Detection and Classification by Decision-Level Fusion for Intelligent Vehicle Systems

PubMed Central

Oh, Sang-Il; Kang, Hang-Bong

2017-01-01

To understand driving environments effectively, it is important to achieve accurate detection and classification of objects detected by sensor-based intelligent vehicle systems, which are significantly important tasks. Object detection is performed for the localization of objects, whereas object classification recognizes object classes from detected object regions. For accurate object detection and classification, fusing multiple sensor information into a key component of the representation and perception processes is necessary. In this paper, we propose a new object-detection and classification method using decision-level fusion. We fuse the classification outputs from independent unary classifiers, such as 3D point clouds and image data using a convolutional neural network (CNN). The unary classifiers for the two sensors are the CNN with five layers, which use more than two pre-trained convolutional layers to consider local to global features as data representation. To represent data using convolutional layers, we apply region of interest (ROI) pooling to the outputs of each layer on the object candidate regions generated using object proposal generation to realize color flattening and semantic grouping for charge-coupled device and Light Detection And Ranging (LiDAR) sensors. We evaluate our proposed method on a KITTI benchmark dataset to detect and classify three object classes: cars, pedestrians and cyclists. The evaluation results show that the proposed method achieves better performance than the previous methods. Our proposed method extracted approximately 500 proposals on a 1226×370 image, whereas the original selective search method extracted approximately 106×n proposals. We obtained classification performance with 77.72% mean average precision over the entirety of the classes in the moderate detection level of the KITTI benchmark dataset. PMID:28117742

Classification of visual and linguistic tasks using eye-movement features.

PubMed

Coco, Moreno I; Keller, Frank

2014-03-07

The role of the task has received special attention in visual-cognition research because it can provide causal explanations of goal-directed eye-movement responses. The dependency between visual attention and task suggests that eye movements can be used to classify the task being performed. A recent study by Greene, Liu, and Wolfe (2012), however, fails to achieve accurate classification of visual tasks based on eye-movement features. In the present study, we hypothesize that tasks can be successfully classified when they differ with respect to the involvement of other cognitive domains, such as language processing. We extract the eye-movement features used by Greene et al. as well as additional features from the data of three different tasks: visual search, object naming, and scene description. First, we demonstrated that eye-movement responses make it possible to characterize the goals of these tasks. Then, we trained three different types of classifiers and predicted the task participants performed with an accuracy well above chance (a maximum of 88% for visual search). An analysis of the relative importance of features for classification accuracy reveals that just one feature, i.e., initiation time, is sufficient for above-chance performance (a maximum of 79% accuracy in object naming). Crucially, this feature is independent of task duration, which differs systematically across the three tasks we investigated. Overall, the best task classification performance was obtained with a set of seven features that included both spatial information (e.g., entropy of attention allocation) and temporal components (e.g., total fixation on objects) of the eye-movement record. This result confirms the task-dependent allocation of visual attention and extends previous work by showing that task classification is possible when tasks differ in the cognitive processes involved (purely visual tasks such as search vs. communicative tasks such as scene description).

Deep convolutional networks for pancreas segmentation in CT imaging

NASA Astrophysics Data System (ADS)

Roth, Holger R.; Farag, Amal; Lu, Le; Turkbey, Evrim B.; Summers, Ronald M.

2015-03-01

Automatic organ segmentation is an important prerequisite for many computer-aided diagnosis systems. The high anatomical variability of organs in the abdomen, such as the pancreas, prevents many segmentation methods from achieving high accuracies when compared to state-of-the-art segmentation of organs like the liver, heart or kidneys. Recently, the availability of large annotated training sets and the accessibility of affordable parallel computing resources via GPUs have made it feasible for "deep learning" methods such as convolutional networks (ConvNets) to succeed in image classification tasks. These methods have the advantage that used classification features are trained directly from the imaging data. We present a fully-automated bottom-up method for pancreas segmentation in computed tomography (CT) images of the abdomen. The method is based on hierarchical coarse-to-fine classification of local image regions (superpixels). Superpixels are extracted from the abdominal region using Simple Linear Iterative Clustering (SLIC). An initial probability response map is generated, using patch-level confidences and a two-level cascade of random forest classifiers, from which superpixel regions with probabilities larger 0.5 are retained. These retained superpixels serve as a highly sensitive initial input of the pancreas and its surroundings to a ConvNet that samples a bounding box around each superpixel at different scales (and random non-rigid deformations at training time) in order to assign a more distinct probability of each superpixel region being pancreas or not. We evaluate our method on CT images of 82 patients (60 for training, 2 for validation, and 20 for testing). Using ConvNets we achieve maximum Dice scores of an average 68% +/- 10% (range, 43-80%) in testing. This shows promise for accurate pancreas segmentation, using a deep learning approach and compares favorably to state-of-the-art methods.

Fast and Robust Segmentation and Classification for Change Detection in Urban Point Clouds

NASA Astrophysics Data System (ADS)

Roynard, X.; Deschaud, J.-E.; Goulette, F.

2016-06-01

Change detection is an important issue in city monitoring to analyse street furniture, road works, car parking, etc. For example, parking surveys are needed but are currently a laborious task involving sending operators in the streets to identify the changes in car locations. In this paper, we propose a method that performs a fast and robust segmentation and classification of urban point clouds, that can be used for change detection. We apply this method to detect the cars, as a particular object class, in order to perform parking surveys automatically. A recently proposed method already addresses the need for fast segmentation and classification of urban point clouds, using elevation images. The interest to work on images is that processing is much faster, proven and robust. However there may be a loss of information in complex 3D cases: for example when objects are one above the other, typically a car under a tree or a pedestrian under a balcony. In this paper we propose a method that retain the three-dimensional information while preserving fast computation times and improving segmentation and classification accuracy. It is based on fast region-growing using an octree, for the segmentation, and specific descriptors with Random-Forest for the classification. Experiments have been performed on large urban point clouds acquired by Mobile Laser Scanning. They show that the method is as fast as the state of the art, and that it gives more robust results in the complex 3D cases.

Free classification of regional dialects of American English

PubMed Central

Clopper, Cynthia G.; Pisoni, David B.

2011-01-01

Recent studies have found that naïve listeners perform poorly in forced-choice dialect categorization tasks. However, the listeners' error patterns in these tasks reveal systematic confusions between phonologically similar dialects. In the present study, a free classification procedure was used to measure the perceptual similarity structure of regional dialect variation in the United States. In two experiments, participants listened to a set of short English sentences produced by male talkers only (Experiment 1) and by male and female talkers (Experiment 2). The listeners were instructed to group the talkers by regional dialect into as many groups as they wanted with as many talkers in each group as they wished. Multidimensional scaling analyses of the data revealed three primary dimensions of perceptual similarity (linguistic markedness, geography, and gender). In addition, a comparison of the results obtained from the free classification task to previous results using the same stimulus materials in six-alternative forced-choice categorization tasks revealed that response biases in the six-alternative task were reduced or eliminated in the free classification task. Thus, the results obtained with the free classification task in the current study provided further evidence that the underlying structure of perceptual dialect category representations reflects important linguistic and sociolinguistic factors. PMID:21423862

Towards a ternary NIRS-BCI: single-trial classification of verbal fluency task, Stroop task and unconstrained rest

NASA Astrophysics Data System (ADS)

Schudlo, Larissa C.; Chau, Tom

2015-12-01

Objective. The majority of near-infrared spectroscopy (NIRS) brain-computer interface (BCI) studies have investigated binary classification problems. Limited work has considered differentiation of more than two mental states, or multi-class differentiation of higher-level cognitive tasks using measurements outside of the anterior prefrontal cortex. Improvements in accuracies are needed to deliver effective communication with a multi-class NIRS system. We investigated the feasibility of a ternary NIRS-BCI that supports mental states corresponding to verbal fluency task (VFT) performance, Stroop task performance, and unconstrained rest using prefrontal and parietal measurements. Approach. Prefrontal and parietal NIRS signals were acquired from 11 able-bodied adults during rest and performance of the VFT or Stroop task. Classification was performed offline using bagging with a linear discriminant base classifier trained on a 10 dimensional feature set. Main results. VFT, Stroop task and rest were classified at an average accuracy of 71.7% ± 7.9%. The ternary classification system provided a statistically significant improvement in information transfer rate relative to a binary system controlled by either mental task (0.87 ± 0.35 bits/min versus 0.73 ± 0.24 bits/min). Significance. These results suggest that effective communication can be achieved with a ternary NIRS-BCI that supports VFT, Stroop task and rest via measurements from the frontal and parietal cortices. Further development of such a system is warranted. Accurate ternary classification can enhance communication rates offered by NIRS-BCIs, improving the practicality of this technology.

Classification of cancerous cells based on the one-class problem approach

NASA Astrophysics Data System (ADS)

Murshed, Nabeel A.; Bortolozzi, Flavio; Sabourin, Robert

1996-03-01

One of the most important factors in reducing the effect of cancerous diseases is the early diagnosis, which requires a good and a robust method. With the advancement of computer technologies and digital image processing, the development of a computer-based system has become feasible. In this paper, we introduce a new approach for the detection of cancerous cells. This approach is based on the one-class problem approach, through which the classification system need only be trained with patterns of cancerous cells. This reduces the burden of the training task by about 50%. Based on this approach, a computer-based classification system is developed, based on the Fuzzy ARTMAP neural networks. Experimental results were performed using a set of 542 patterns taken from a sample of breast cancer. Results of the experiment show 98% correct identification of cancerous cells and 95% correct identification of non-cancerous cells.

Patient-Specific Deep Architectural Model for ECG Classification

PubMed Central

Luo, Kan; Cuschieri, Alfred

2017-01-01

Heartbeat classification is a crucial step for arrhythmia diagnosis during electrocardiographic (ECG) analysis. The new scenario of wireless body sensor network- (WBSN-) enabled ECG monitoring puts forward a higher-level demand for this traditional ECG analysis task. Previously reported methods mainly addressed this requirement with the applications of a shallow structured classifier and expert-designed features. In this study, modified frequency slice wavelet transform (MFSWT) was firstly employed to produce the time-frequency image for heartbeat signal. Then the deep learning (DL) method was performed for the heartbeat classification. Here, we proposed a novel model incorporating automatic feature abstraction and a deep neural network (DNN) classifier. Features were automatically abstracted by the stacked denoising auto-encoder (SDA) from the transferred time-frequency image. DNN classifier was constructed by an encoder layer of SDA and a softmax layer. In addition, a deterministic patient-specific heartbeat classifier was achieved by fine-tuning on heartbeat samples, which included a small subset of individual samples. The performance of the proposed model was evaluated on the MIT-BIH arrhythmia database. Results showed that an overall accuracy of 97.5% was achieved using the proposed model, confirming that the proposed DNN model is a powerful tool for heartbeat pattern recognition. PMID:29065597

Multimedia Classifier

NASA Astrophysics Data System (ADS)

Costache, G. N.; Gavat, I.

2004-09-01

Along with the aggressive growing of the amount of digital data available (text, audio samples, digital photos and digital movies joined all in the multimedia domain) the need for classification, recognition and retrieval of this kind of data became very important. In this paper will be presented a system structure to handle multimedia data based on a recognition perspective. The main processing steps realized for the interesting multimedia objects are: first, the parameterization, by analysis, in order to obtain a description based on features, forming the parameter vector; second, a classification, generally with a hierarchical structure to make the necessary decisions. For audio signals, both speech and music, the derived perceptual features are the melcepstral (MFCC) and the perceptual linear predictive (PLP) coefficients. For images, the derived features are the geometric parameters of the speaker mouth. The hierarchical classifier consists generally in a clustering stage, based on the Kohonnen Self-Organizing Maps (SOM) and a final stage, based on a powerful classification algorithm called Support Vector Machines (SVM). The system, in specific variants, is applied with good results in two tasks: the first, is a bimodal speech recognition which uses features obtained from speech signal fused to features obtained from speaker's image and the second is a music retrieval from large music database.

SAR image classification based on CNN in real and simulation datasets

NASA Astrophysics Data System (ADS)

Peng, Lijiang; Liu, Ming; Liu, Xiaohua; Dong, Liquan; Hui, Mei; Zhao, Yuejin

2018-04-01

Convolution neural network (CNN) has made great success in image classification tasks. Even in the field of synthetic aperture radar automatic target recognition (SAR-ATR), state-of-art results has been obtained by learning deep representation of features on the MSTAR benchmark. However, the raw data of MSTAR have shortcomings in training a SAR-ATR model because of high similarity in background among the SAR images of each kind. This indicates that the CNN would learn the hierarchies of features of backgrounds as well as the targets. To validate the influence of the background, some other SAR images datasets have been made which contains the simulation SAR images of 10 manufactured targets such as tank and fighter aircraft, and the backgrounds of simulation SAR images are sampled from the whole original MSTAR data. The simulation datasets contain the dataset that the backgrounds of each kind images correspond to the one kind of backgrounds of MSTAR targets or clutters and the dataset that each image shares the random background of whole MSTAR targets or clutters. In addition, mixed datasets of MSTAR and simulation datasets had been made to use in the experiments. The CNN architecture proposed in this paper are trained on all datasets mentioned above. The experimental results shows that the architecture can get high performances on all datasets even the backgrounds of the images are miscellaneous, which indicates the architecture can learn a good representation of the targets even though the drastic changes on background.

A fast image retrieval method based on SVM and imbalanced samples in filtering multimedia message spam

NASA Astrophysics Data System (ADS)

Chen, Zhang; Peng, Zhenming; Peng, Lingbing; Liao, Dongyi; He, Xin

2011-11-01

With the swift and violent development of the Multimedia Messaging Service (MMS), it becomes an urgent task to filter the Multimedia Message (MM) spam effectively in real-time. For the fact that most MMs contain images or videos, a method based on retrieving images is given in this paper for filtering MM spam. The detection method used in this paper is a combination of skin-color detection, texture detection, and face detection, and the classifier for this imbalanced problem is a very fast multi-classification combining Support vector machine (SVM) with unilateral binary decision tree. The experiments on 3 test sets show that the proposed method is effective, with the interception rate up to 60% and the average detection time for each image less than 1 second.

Integrating dimension reduction and out-of-sample extension in automated classification of ex vivo human patellar cartilage on phase contrast X-ray computed tomography.

PubMed

Nagarajan, Mahesh B; Coan, Paola; Huber, Markus B; Diemoz, Paul C; Wismüller, Axel

2015-01-01

Phase contrast X-ray computed tomography (PCI-CT) has been demonstrated as a novel imaging technique that can visualize human cartilage with high spatial resolution and soft tissue contrast. Different textural approaches have been previously investigated for characterizing chondrocyte organization on PCI-CT to enable classification of healthy and osteoarthritic cartilage. However, the large size of feature sets extracted in such studies motivates an investigation into algorithmic feature reduction for computing efficient feature representations without compromising their discriminatory power. For this purpose, geometrical feature sets derived from the scaling index method (SIM) were extracted from 1392 volumes of interest (VOI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. The extracted feature sets were subject to linear and non-linear dimension reduction techniques as well as feature selection based on evaluation of mutual information criteria. The reduced feature set was subsequently used in a machine learning task with support vector regression to classify VOIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver-operating characteristic (ROC) curve (AUC). Our results show that the classification performance achieved by 9-D SIM-derived geometric feature sets (AUC: 0.96 ± 0.02) can be maintained with 2-D representations computed from both dimension reduction and feature selection (AUC values as high as 0.97 ± 0.02). Thus, such feature reduction techniques can offer a high degree of compaction to large feature sets extracted from PCI-CT images while maintaining their ability to characterize the underlying chondrocyte patterns.

Performance of an Artificial Multi-observer Deep Neural Network for Fully Automated Segmentation of Polycystic Kidneys.

PubMed

Kline, Timothy L; Korfiatis, Panagiotis; Edwards, Marie E; Blais, Jaime D; Czerwiec, Frank S; Harris, Peter C; King, Bernard F; Torres, Vicente E; Erickson, Bradley J

2017-08-01

Deep learning techniques are being rapidly applied to medical imaging tasks-from organ and lesion segmentation to tissue and tumor classification. These techniques are becoming the leading algorithmic approaches to solve inherently difficult image processing tasks. Currently, the most critical requirement for successful implementation lies in the need for relatively large datasets that can be used for training the deep learning networks. Based on our initial studies of MR imaging examinations of the kidneys of patients affected by polycystic kidney disease (PKD), we have generated a unique database of imaging data and corresponding reference standard segmentations of polycystic kidneys. In the study of PKD, segmentation of the kidneys is needed in order to measure total kidney volume (TKV). Automated methods to segment the kidneys and measure TKV are needed to increase measurement throughput and alleviate the inherent variability of human-derived measurements. We hypothesize that deep learning techniques can be leveraged to perform fast, accurate, reproducible, and fully automated segmentation of polycystic kidneys. Here, we describe a fully automated approach for segmenting PKD kidneys within MR images that simulates a multi-observer approach in order to create an accurate and robust method for the task of segmentation and computation of TKV for PKD patients. A total of 2000 cases were used for training and validation, and 400 cases were used for testing. The multi-observer ensemble method had mean ± SD percent volume difference of 0.68 ± 2.2% compared with the reference standard segmentations. The complete framework performs fully automated segmentation at a level comparable with interobserver variability and could be considered as a replacement for the task of segmentation of PKD kidneys by a human.

Volumetric characterization of human patellar cartilage matrix on phase contrast x-ray computed tomography

NASA Astrophysics Data System (ADS)

Abidin, Anas Z.; Nagarajan, Mahesh B.; Checefsky, Walter A.; Coan, Paola; Diemoz, Paul C.; Hobbs, Susan K.; Huber, Markus B.; Wismüller, Axel

2015-03-01

Phase contrast X-ray computed tomography (PCI-CT) has recently emerged as a novel imaging technique that allows visualization of cartilage soft tissue, subsequent examination of chondrocyte patterns, and their correlation to osteoarthritis. Previous studies have shown that 2D texture features are effective at distinguishing between healthy and osteoarthritic regions of interest annotated in the radial zone of cartilage matrix on PCI-CT images. In this study, we further extend the texture analysis to 3D and investigate the ability of volumetric texture features at characterizing chondrocyte patterns in the cartilage matrix for purposes of classification. Here, we extracted volumetric texture features derived from Minkowski Functionals and gray-level co-occurrence matrices (GLCM) from 496 volumes of interest (VOI) annotated on PCI-CT images of human patellar cartilage specimens. The extracted features were then used in a machine-learning task involving support vector regression to classify ROIs as healthy or osteoarthritic. Classification performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). The best classification performance was observed with GLCM features correlation (AUC = 0.83 +/- 0.06) and homogeneity (AUC = 0.82 +/- 0.07), which significantly outperformed all Minkowski Functionals (p < 0.05). These results suggest that such quantitative analysis of chondrocyte patterns in human patellar cartilage matrix involving GLCM-derived statistical features can distinguish between healthy and osteoarthritic tissue with high accuracy.

Classification and recognition of dynamical models: the role of phase, independent components, kernels and optimal transport.

PubMed

Bissacco, Alessandro; Chiuso, Alessandro; Soatto, Stefano

2007-11-01

We address the problem of performing decision tasks, and in particular classification and recognition, in the space of dynamical models in order to compare time series of data. Motivated by the application of recognition of human motion in image sequences, we consider a class of models that include linear dynamics, both stable and marginally stable (periodic), both minimum and non-minimum phase, driven by non-Gaussian processes. This requires extending existing learning and system identification algorithms to handle periodic modes and nonminimum phase behavior, while taking into account higher-order statistics of the data. Once a model is identified, we define a kernel-based cord distance between models that includes their dynamics, their initial conditions as well as input distribution. This is made possible by a novel kernel defined between two arbitrary (non-Gaussian) distributions, which is computed by efficiently solving an optimal transport problem. We validate our choice of models, inference algorithm, and distance on the tasks of human motion synthesis (sample paths of the learned models), and recognition (nearest-neighbor classification in the computed distance). However, our work can be applied more broadly where one needs to compare historical data while taking into account periodic trends, non-minimum phase behavior, and non-Gaussian input distributions.

Weakly Supervised Dictionary Learning

NASA Astrophysics Data System (ADS)

You, Zeyu; Raich, Raviv; Fern, Xiaoli Z.; Kim, Jinsub

2018-05-01

We present a probabilistic modeling and inference framework for discriminative analysis dictionary learning under a weak supervision setting. Dictionary learning approaches have been widely used for tasks such as low-level signal denoising and restoration as well as high-level classification tasks, which can be applied to audio and image analysis. Synthesis dictionary learning aims at jointly learning a dictionary and corresponding sparse coefficients to provide accurate data representation. This approach is useful for denoising and signal restoration, but may lead to sub-optimal classification performance. By contrast, analysis dictionary learning provides a transform that maps data to a sparse discriminative representation suitable for classification. We consider the problem of analysis dictionary learning for time-series data under a weak supervision setting in which signals are assigned with a global label instead of an instantaneous label signal. We propose a discriminative probabilistic model that incorporates both label information and sparsity constraints on the underlying latent instantaneous label signal using cardinality control. We present the expectation maximization (EM) procedure for maximum likelihood estimation (MLE) of the proposed model. To facilitate a computationally efficient E-step, we propose both a chain and a novel tree graph reformulation of the graphical model. The performance of the proposed model is demonstrated on both synthetic and real-world data.

Advancing Bag-of-Visual-Words Representations for Lesion Classification in Retinal Images

PubMed Central

Pires, Ramon; Jelinek, Herbert F.; Wainer, Jacques; Valle, Eduardo; Rocha, Anderson

2014-01-01

Diabetic Retinopathy (DR) is a complication of diabetes that can lead to blindness if not readily discovered. Automated screening algorithms have the potential to improve identification of patients who need further medical attention. However, the identification of lesions must be accurate to be useful for clinical application. The bag-of-visual-words (BoVW) algorithm employs a maximum-margin classifier in a flexible framework that is able to detect the most common DR-related lesions such as microaneurysms, cotton-wool spots and hard exudates. BoVW allows to bypass the need for pre- and post-processing of the retinographic images, as well as the need of specific ad hoc techniques for identification of each type of lesion. An extensive evaluation of the BoVW model, using three large retinograph datasets (DR1, DR2 and Messidor) with different resolution and collected by different healthcare personnel, was performed. The results demonstrate that the BoVW classification approach can identify different lesions within an image without having to utilize different algorithms for each lesion reducing processing time and providing a more flexible diagnostic system. Our BoVW scheme is based on sparse low-level feature detection with a Speeded-Up Robust Features (SURF) local descriptor, and mid-level features based on semi-soft coding with max pooling. The best BoVW representation for retinal image classification was an area under the receiver operating characteristic curve (AUC-ROC) of 97.8% (exudates) and 93.5% (red lesions), applying a cross-dataset validation protocol. To assess the accuracy for detecting cases that require referral within one year, the sparse extraction technique associated with semi-soft coding and max pooling obtained an AUC of 94.22.0%, outperforming current methods. Those results indicate that, for retinal image classification tasks in clinical practice, BoVW is equal and, in some instances, surpasses results obtained using dense detection (widely believed to be the best choice in many vision problems) for the low-level descriptors. PMID:24886780

Some new classification methods for hyperspectral remote sensing

NASA Astrophysics Data System (ADS)

Du, Pei-jun; Chen, Yun-hao; Jones, Simon; Ferwerda, Jelle G.; Chen, Zhi-jun; Zhang, Hua-peng; Tan, Kun; Yin, Zuo-xia

2006-10-01

Hyperspectral Remote Sensing (HRS) is one of the most significant recent achievements of Earth Observation Technology. Classification is the most commonly employed processing methodology. In this paper three new hyperspectral RS image classification methods are analyzed. These methods are: Object-oriented FIRS image classification, HRS image classification based on information fusion and HSRS image classification by Back Propagation Neural Network (BPNN). OMIS FIRS image is used as the example data. Object-oriented techniques have gained popularity for RS image classification in recent years. In such method, image segmentation is used to extract the regions from the pixel information based on homogeneity criteria at first, and spectral parameters like mean vector, texture, NDVI and spatial/shape parameters like aspect ratio, convexity, solidity, roundness and orientation for each region are calculated, finally classification of the image using the region feature vectors and also using suitable classifiers such as artificial neural network (ANN). It proves that object-oriented methods can improve classification accuracy since they utilize information and features both from the point and the neighborhood, and the processing unit is a polygon (in which all pixels are homogeneous and belong to the class). HRS image classification based on information fusion, divides all bands of the image into different groups initially, and extracts features from every group according to the properties of each group. Three levels of information fusion: data level fusion, feature level fusion and decision level fusion are used to HRS image classification. Artificial Neural Network (ANN) can perform well in RS image classification. In order to promote the advances of ANN used for HIRS image classification, Back Propagation Neural Network (BPNN), the most commonly used neural network, is used to HRS image classification.

Detection of Pigment Networks in Dermoscopy Images

NASA Astrophysics Data System (ADS)

Eltayef, Khalid; Li, Yongmin; Liu, Xiaohui

2017-02-01

One of the most important structures in dermoscopy images is the pigment network, which is also one of the most challenging and fundamental task for dermatologists in early detection of melanoma. This paper presents an automatic system to detect pigment network from dermoscopy images. The design of the proposed algorithm consists of four stages. First, a pre-processing algorithm is carried out in order to remove the noise and improve the quality of the image. Second, a bank of directional filters and morphological connected component analysis are applied to detect the pigment networks. Third, features are extracted from the detected image, which can be used in the subsequent stage. Fourth, the classification process is performed by applying feed-forward neural network, in order to classify the region as either normal or abnormal skin. The method was tested on a dataset of 200 dermoscopy images from Hospital Pedro Hispano (Matosinhos), and better results were produced compared to previous studies.

Towards a computer-aided diagnosis system for vocal cord diseases.

PubMed

Verikas, A; Gelzinis, A; Bacauskiene, M; Uloza, V

2006-01-01

The objective of this work is to investigate a possibility of creating a computer-aided decision support system for an automated analysis of vocal cord images aiming to categorize diseases of vocal cords. The problem is treated as a pattern recognition task. To obtain a concise and informative representation of a vocal cord image, colour, texture, and geometrical features are used. The representation is further analyzed by a pattern classifier categorizing the image into healthy, diffuse, and nodular classes. The approach developed was tested on 785 vocal cord images collected at the Department of Otolaryngology, Kaunas University of Medicine, Lithuania. A correct classification rate of over 87% was obtained when categorizing a set of unseen images into the aforementioned three classes. Bearing in mind the high similarity of the decision classes, the results obtained are rather encouraging and the developed tools could be very helpful for assuring objective analysis of the images of laryngeal diseases.

Research on Remote Sensing Image Classification Based on Feature Level Fusion

NASA Astrophysics Data System (ADS)

Yuan, L.; Zhu, G.

2018-04-01

Remote sensing image classification, as an important direction of remote sensing image processing and application, has been widely studied. However, in the process of existing classification algorithms, there still exists the phenomenon of misclassification and missing points, which leads to the final classification accuracy is not high. In this paper, we selected Sentinel-1A and Landsat8 OLI images as data sources, and propose a classification method based on feature level fusion. Compare three kind of feature level fusion algorithms (i.e., Gram-Schmidt spectral sharpening, Principal Component Analysis transform and Brovey transform), and then select the best fused image for the classification experimental. In the classification process, we choose four kinds of image classification algorithms (i.e. Minimum distance, Mahalanobis distance, Support Vector Machine and ISODATA) to do contrast experiment. We use overall classification precision and Kappa coefficient as the classification accuracy evaluation criteria, and the four classification results of fused image are analysed. The experimental results show that the fusion effect of Gram-Schmidt spectral sharpening is better than other methods. In four kinds of classification algorithms, the fused image has the best applicability to Support Vector Machine classification, the overall classification precision is 94.01 % and the Kappa coefficients is 0.91. The fused image with Sentinel-1A and Landsat8 OLI is not only have more spatial information and spectral texture characteristics, but also enhances the distinguishing features of the images. The proposed method is beneficial to improve the accuracy and stability of remote sensing image classification.

Exploring Models and Data for Remote Sensing Image Caption Generation

NASA Astrophysics Data System (ADS)

Lu, Xiaoqiang; Wang, Binqiang; Zheng, Xiangtao; Li, Xuelong

2018-04-01

Inspired by recent development of artificial satellite, remote sensing images have attracted extensive attention. Recently, noticeable progress has been made in scene classification and target detection.However, it is still not clear how to describe the remote sensing image content with accurate and concise sentences. In this paper, we investigate to describe the remote sensing images with accurate and flexible sentences. First, some annotated instructions are presented to better describe the remote sensing images considering the special characteristics of remote sensing images. Second, in order to exhaustively exploit the contents of remote sensing images, a large-scale aerial image data set is constructed for remote sensing image caption. Finally, a comprehensive review is presented on the proposed data set to fully advance the task of remote sensing caption. Extensive experiments on the proposed data set demonstrate that the content of the remote sensing image can be completely described by generating language descriptions. The data set is available at https://github.com/201528014227051/RSICD_optimal

Development of classification models to detect Salmonella Enteritidis and Salmonella Typhimurium found in poultry carcass rinses by visible-near infrared hyperspectral imaging

NASA Astrophysics Data System (ADS)

Seo, Young Wook; Yoon, Seung Chul; Park, Bosoon; Hinton, Arthur; Windham, William R.; Lawrence, Kurt C.

2013-05-01

Salmonella is a major cause of foodborne disease outbreaks resulting from the consumption of contaminated food products in the United States. This paper reports the development of a hyperspectral imaging technique for detecting and differentiating two of the most common Salmonella serotypes, Salmonella Enteritidis (SE) and Salmonella Typhimurium (ST), from background microflora that are often found in poultry carcass rinse. Presumptive positive screening of colonies with a traditional direct plating method is a labor intensive and time consuming task. Thus, this paper is concerned with the detection of differences in spectral characteristics among the pure SE, ST, and background microflora grown on brilliant green sulfa (BGS) and xylose lysine tergitol 4 (XLT4) agar media with a spread plating technique. Visible near-infrared hyperspectral imaging, providing the spectral and spatial information unique to each microorganism, was utilized to differentiate SE and ST from the background microflora. A total of 10 classification models, including five machine learning algorithms, each without and with principal component analysis (PCA), were validated and compared to find the best model in classification accuracy. The five machine learning (classification) algorithms used in this study were Mahalanobis distance (MD), k-nearest neighbor (kNN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and support vector machine (SVM). The average classification accuracy of all 10 models on a calibration (or training) set of the pure cultures on BGS agar plates was 98% (Kappa coefficient = 0.95) in determining the presence of SE and/or ST although it was difficult to differentiate between SE and ST. The average classification accuracy of all 10 models on a training set for ST detection on XLT4 agar was over 99% (Kappa coefficient = 0.99) although SE colonies on XLT4 agar were difficult to differentiate from background microflora. The average classification accuracy of all 10 models on a validation set of chicken carcass rinses spiked with SE or ST and incubated on BGS agar plates was 94.45% and 83.73%, without and with PCA for classification, respectively. The best performing classification model on the validation set was QDA without PCA by achieving the classification accuracy of 98.65% (Kappa coefficient=0.98). The overall best performing classification model regardless of using PCA was MD with the classification accuracy of 94.84% (Kappa coefficient=0.88) on the validation set.

Automatic Sea Bird Detection from High Resolution Aerial Imagery

NASA Astrophysics Data System (ADS)

Mader, S.; Grenzdörffer, G. J.

2016-06-01

Great efforts are presently taken in the scientific community to develop computerized and (fully) automated image processing methods allowing for an efficient and automatic monitoring of sea birds and marine mammals in ever-growing amounts of aerial imagery. Currently the major part of the processing, however, is still conducted by especially trained professionals, visually examining the images and detecting and classifying the requested subjects. This is a very tedious task, particularly when the rate of void images regularly exceeds the mark of 90%. In the content of this contribution we will present our work aiming to support the processing of aerial images by modern methods from the field of image processing. We will especially focus on the combination of local, region-based feature detection and piecewise global image segmentation for automatic detection of different sea bird species. Large image dimensions resulting from the use of medium and large-format digital cameras in aerial surveys inhibit the applicability of image processing methods based on global operations. In order to efficiently handle those image sizes and to nevertheless take advantage of globally operating segmentation algorithms, we will describe the combined usage of a simple performant feature detector based on local operations on the original image with a complex global segmentation algorithm operating on extracted sub-images. The resulting exact segmentation of possible candidates then serves as a basis for the determination of feature vectors for subsequent elimination of false candidates and for classification tasks.

Texture for script identification.

PubMed

Busch, Andrew; Boles, Wageeh W; Sridharan, Sridha

2005-11-01

The problem of determining the script and language of a document image has a number of important applications in the field of document analysis, such as indexing and sorting of large collections of such images, or as a precursor to optical character recognition (OCR). In this paper, we investigate the use of texture as a tool for determining the script of a document image, based on the observation that text has a distinct visual texture. An experimental evaluation of a number of commonly used texture features is conducted on a newly created script database, providing a qualitative measure of which features are most appropriate for this task. Strategies for improving classification results in situations with limited training data and multiple font types are also proposed.

Infrared small target detection with kernel Fukunaga Koontz transform

NASA Astrophysics Data System (ADS)

Liu, Rui-ming; Liu, Er-qi; Yang, Jie; Zhang, Tian-hao; Wang, Fang-lin

2007-09-01

The Fukunaga-Koontz transform (FKT) has been proposed for many years. It can be used to solve two-pattern classification problems successfully. However, there are few researchers who have definitely extended FKT to kernel FKT (KFKT). In this paper, we first complete this task. Then a method based on KFKT is developed to detect infrared small targets. KFKT is a supervised learning algorithm. How to construct training sets is very important. For automatically detecting targets, the synthetic target images and real background images are used to train KFKT. Because KFKT can represent the higher order statistical properties of images, we expect better detection performance of KFKT than that of FKT. The well-devised experiments verify that KFKT outperforms FKT in detecting infrared small targets.

NutriNet: A Deep Learning Food and Drink Image Recognition System for Dietary Assessment

PubMed Central

Koroušić Seljak, Barbara

2017-01-01

Automatic food image recognition systems are alleviating the process of food-intake estimation and dietary assessment. However, due to the nature of food images, their recognition is a particularly challenging task, which is why traditional approaches in the field have achieved a low classification accuracy. Deep neural networks have outperformed such solutions, and we present a novel approach to the problem of food and drink image detection and recognition that uses a newly-defined deep convolutional neural network architecture, called NutriNet. This architecture was tuned on a recognition dataset containing 225,953 512 × 512 pixel images of 520 different food and drink items from a broad spectrum of food groups, on which we achieved a classification accuracy of 86.72%, along with an accuracy of 94.47% on a detection dataset containing 130,517 images. We also performed a real-world test on a dataset of self-acquired images, combined with images from Parkinson’s disease patients, all taken using a smartphone camera, achieving a top-five accuracy of 55%, which is an encouraging result for real-world images. Additionally, we tested NutriNet on the University of Milano-Bicocca 2016 (UNIMIB2016) food image dataset, on which we improved upon the provided baseline recognition result. An online training component was implemented to continually fine-tune the food and drink recognition model on new images. The model is being used in practice as part of a mobile app for the dietary assessment of Parkinson’s disease patients. PMID:28653995

Modelling the influence of noise of the image sensor for blood cells recognition in computer microscopy

NASA Astrophysics Data System (ADS)

Nikitaev, V. G.; Nagornov, O. V.; Pronichev, A. N.; Polyakov, E. V.; Dmitrieva, V. V.

2017-12-01

The first stage of diagnostics of blood cancer is the analysis of blood smears. The application of decision-making support systems would reduce the subjectivity of the diagnostic process and avoid errors, resulting in often irreversible changes in the patient's condition. In this regard, the solution of this problem requires the use of modern technology. One of the tools of the program classification of blood cells are texture features, and the task of finding informative among them is promising. The paper investigates the effect of noise of the image sensor to informative texture features with application of methods of mathematical modelling.

A generic nuclei detection method for histopathological breast images

NASA Astrophysics Data System (ADS)

Kost, Henning; Homeyer, André; Bult, Peter; Balkenhol, Maschenka C. A.; van der Laak, Jeroen A. W. M.; Hahn, Horst K.

2016-03-01

The detection of cell nuclei plays a key role in various histopathological image analysis problems. Considering the high variability of its applications, we propose a novel generic and trainable detection approach. Adaption to specific nuclei detection tasks is done by providing training samples. A trainable deconvolution and classification algorithm is used to generate a probability map indicating the presence of a nucleus. The map is processed by an extended watershed segmentation step to identify the nuclei positions. We have tested our method on data sets with different stains and target nuclear types. We obtained F1-measures between 0.83 and 0.93.

Classification of galaxy type from images using Microsoft R Server

NASA Astrophysics Data System (ADS)

de Vries, Andrie

2017-06-01

Many astronomers working in the field of AstroInformatics write code as part of their work. Although the programming language of choice is Python, a small number (8%) use R. R has its specific strengths in the domain of statistics, and is often viewed as limited in the size of data it can handle. However, Microsoft R Server is a product that removes these limitations by being able to process much larger amounts of data. I present some highlights of R Server, by illustrating how to fit a convolutional neural network using R. The specific task is to classify galaxies, using only images extracted from the Sloan Digital Skyserver.

The interplay of holistic shape, local feature and color information in object categorization.

PubMed

Rokszin, Adrienn Aranka; Győri-Dani, Dóra; Linnert, Szilvia; Krajcsi, Attila; Tompa, Tamás; Csifcsák, Gábor

2015-07-01

Although it is widely accepted that colors facilitate object and scene recognition under various circumstances, several studies found no effects of color removal in tasks requiring categorization of briefly presented animals in natural scenes. In this study, three experiments were performed to test the assumption that the discrepancy between empirical data is related to variations of the available meaningful global information such as object shapes and contextual cues. Sixty-one individuals categorized chromatic and achromatic versions of intact and scrambled images containing either cars or birds. While color removal did not affect the classification of intact stimuli, the recognition of moderately scrambled achromatic images was more difficult. This effect was accompanied by amplitude modulations of occipital event-related potentials emerging from approximately 150ms post-stimulus. Our results indicate that colors facilitate stimulus classification, but this effect becomes prominent only in cases when holistic processing is not sufficient for stimulus recognition. Copyright © 2015 Elsevier B.V. All rights reserved.

How musical expertise shapes speech perception: evidence from auditory classification images.

PubMed

Varnet, Léo; Wang, Tianyun; Peter, Chloe; Meunier, Fanny; Hoen, Michel

2015-09-24

It is now well established that extensive musical training percolates to higher levels of cognition, such as speech processing. However, the lack of a precise technique to investigate the specific listening strategy involved in speech comprehension has made it difficult to determine how musicians' higher performance in non-speech tasks contributes to their enhanced speech comprehension. The recently developed Auditory Classification Image approach reveals the precise time-frequency regions used by participants when performing phonemic categorizations in noise. Here we used this technique on 19 non-musicians and 19 professional musicians. We found that both groups used very similar listening strategies, but the musicians relied more heavily on the two main acoustic cues, at the first formant onset and at the onsets of the second and third formants onsets. Additionally, they responded more consistently to stimuli. These observations provide a direct visualization of auditory plasticity resulting from extensive musical training and shed light on the level of functional transfer between auditory processing and speech perception.

Multisource Data Classification Using A Hybrid Semi-supervised Learning Scheme

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

Vatsavai, Raju; Bhaduri, Budhendra L; Shekhar, Shashi

2009-01-01

In many practical situations thematic classes can not be discriminated by spectral measurements alone. Often one needs additional features such as population density, road density, wetlands, elevation, soil types, etc. which are discrete attributes. On the other hand remote sensing image features are continuous attributes. Finding a suitable statistical model and estimation of parameters is a challenging task in multisource (e.g., discrete and continuous attributes) data classification. In this paper we present a semi-supervised learning method by assuming that the samples were generated by a mixture model, where each component could be either a continuous or discrete distribution. Overall classificationmore » accuracy of the proposed method is improved by 12% in our initial experiments.« less

Commission 5: Documentation and Astronomical Data

NASA Astrophysics Data System (ADS)

Norris, Raymond P.; Ohishi, Masatoshi; Genova, Françoise; Grothkopf, Uta; Malkov, Oleg Yu.; Pence, William D.; Schmitz, Marion; Hanisch, Robert J.; Zhou, Xu

IAU Commission 5 deals with data management issues, and its working groups and task groups deal specifically with information handling, with data centres and networks, with technical aspects of collection, archiving, storage and dissemination of data, with designations and classification of astronomical objects, with library services, editorial policies, computer communications, ad hoc methodologies, and with various standards, reference frames, etc., FITS, astronomys Flexible Image Transport System, the major data exchange format, is controlled, maintained and updated by the Working Group FITS.

Investigation of automated feature extraction using multiple data sources

NASA Astrophysics Data System (ADS)

Harvey, Neal R.; Perkins, Simon J.; Pope, Paul A.; Theiler, James P.; David, Nancy A.; Porter, Reid B.

2003-04-01

An increasing number and variety of platforms are now capable of collecting remote sensing data over a particular scene. For many applications, the information available from any individual sensor may be incomplete, inconsistent or imprecise. However, other sources may provide complementary and/or additional data. Thus, for an application such as image feature extraction or classification, it may be that fusing the mulitple data sources can lead to more consistent and reliable results. Unfortunately, with the increased complexity of the fused data, the search space of feature-extraction or classification algorithms also greatly increases. With a single data source, the determination of a suitable algorithm may be a significant challenge for an image analyst. With the fused data, the search for suitable algorithms can go far beyond the capabilities of a human in a realistic time frame, and becomes the realm of machine learning, where the computational power of modern computers can be harnessed to the task at hand. We describe experiments in which we investigate the ability of a suite of automated feature extraction tools developed at Los Alamos National Laboratory to make use of multiple data sources for various feature extraction tasks. We compare and contrast this software's capabilities on 1) individual data sets from different data sources 2) fused data sets from multiple data sources and 3) fusion of results from multiple individual data sources.

A deep learning method for classifying mammographic breast density categories.

PubMed

Mohamed, Aly A; Berg, Wendie A; Peng, Hong; Luo, Yahong; Jankowitz, Rachel C; Wu, Shandong

2018-01-01

Mammographic breast density is an established risk marker for breast cancer and is visually assessed by radiologists in routine mammogram image reading, using four qualitative Breast Imaging and Reporting Data System (BI-RADS) breast density categories. It is particularly difficult for radiologists to consistently distinguish the two most common and most variably assigned BI-RADS categories, i.e., "scattered density" and "heterogeneously dense". The aim of this work was to investigate a deep learning-based breast density classifier to consistently distinguish these two categories, aiming at providing a potential computerized tool to assist radiologists in assigning a BI-RADS category in current clinical workflow. In this study, we constructed a convolutional neural network (CNN)-based model coupled with a large (i.e., 22,000 images) digital mammogram imaging dataset to evaluate the classification performance between the two aforementioned breast density categories. All images were collected from a cohort of 1,427 women who underwent standard digital mammography screening from 2005 to 2016 at our institution. The truths of the density categories were based on standard clinical assessment made by board-certified breast imaging radiologists. Effects of direct training from scratch solely using digital mammogram images and transfer learning of a pretrained model on a large nonmedical imaging dataset were evaluated for the specific task of breast density classification. In order to measure the classification performance, the CNN classifier was also tested on a refined version of the mammogram image dataset by removing some potentially inaccurately labeled images. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were used to measure the accuracy of the classifier. The AUC was 0.9421 when the CNN-model was trained from scratch on our own mammogram images, and the accuracy increased gradually along with an increased size of training samples. Using the pretrained model followed by a fine-tuning process with as few as 500 mammogram images led to an AUC of 0.9265. After removing the potentially inaccurately labeled images, AUC was increased to 0.9882 and 0.9857 for without and with the pretrained model, respectively, both significantly higher (P < 0.001) than when using the full imaging dataset. Our study demonstrated high classification accuracies between two difficult to distinguish breast density categories that are routinely assessed by radiologists. We anticipate that our approach will help enhance current clinical assessment of breast density and better support consistent density notification to patients in breast cancer screening. © 2017 American Association of Physicists in Medicine.

LUNGx Challenge for computerized lung nodule classification

DOE PAGES

Armato, Samuel G.; Drukker, Karen; Li, Feng; ...

2016-12-19

The purpose of this work is to describe the LUNGx Challenge for the computerized classification of lung nodules on diagnostic computed tomography (CT) scans as benign or malignant and report the performance of participants’ computerized methods along with that of six radiologists who participated in an observer study performing the same Challenge task on the same dataset. The Challenge provided sets of calibration and testing scans, established a performance assessment process, and created an infrastructure for case dissemination and result submission. We present ten groups that applied their own methods to 73 lung nodules (37 benign and 36 malignant) thatmore » were selected to achieve approximate size matching between the two cohorts. Area under the receiver operating characteristic curve (AUC) values for these methods ranged from 0.50 to 0.68; only three methods performed statistically better than random guessing. The radiologists’ AUC values ranged from 0.70 to 0.85; three radiologists performed statistically better than the best-performing computer method. The LUNGx Challenge compared the performance of computerized methods in the task of differentiating benign from malignant lung nodules on CT scans, placed in the context of the performance of radiologists on the same task. Lastly, the continued public availability of the Challenge cases will provide a valuable resource for the medical imaging research community.« less

LUNGx Challenge for computerized lung nodule classification

PubMed Central

Armato, Samuel G.; Drukker, Karen; Li, Feng; Hadjiiski, Lubomir; Tourassi, Georgia D.; Engelmann, Roger M.; Giger, Maryellen L.; Redmond, George; Farahani, Keyvan; Kirby, Justin S.; Clarke, Laurence P.

2016-01-01

Abstract. The purpose of this work is to describe the LUNGx Challenge for the computerized classification of lung nodules on diagnostic computed tomography (CT) scans as benign or malignant and report the performance of participants’ computerized methods along with that of six radiologists who participated in an observer study performing the same Challenge task on the same dataset. The Challenge provided sets of calibration and testing scans, established a performance assessment process, and created an infrastructure for case dissemination and result submission. Ten groups applied their own methods to 73 lung nodules (37 benign and 36 malignant) that were selected to achieve approximate size matching between the two cohorts. Area under the receiver operating characteristic curve (AUC) values for these methods ranged from 0.50 to 0.68; only three methods performed statistically better than random guessing. The radiologists’ AUC values ranged from 0.70 to 0.85; three radiologists performed statistically better than the best-performing computer method. The LUNGx Challenge compared the performance of computerized methods in the task of differentiating benign from malignant lung nodules on CT scans, placed in the context of the performance of radiologists on the same task. The continued public availability of the Challenge cases will provide a valuable resource for the medical imaging research community. PMID:28018939

LUNGx Challenge for computerized lung nodule classification

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

Armato, Samuel G.; Drukker, Karen; Li, Feng

The purpose of this work is to describe the LUNGx Challenge for the computerized classification of lung nodules on diagnostic computed tomography (CT) scans as benign or malignant and report the performance of participants’ computerized methods along with that of six radiologists who participated in an observer study performing the same Challenge task on the same dataset. The Challenge provided sets of calibration and testing scans, established a performance assessment process, and created an infrastructure for case dissemination and result submission. We present ten groups that applied their own methods to 73 lung nodules (37 benign and 36 malignant) thatmore » were selected to achieve approximate size matching between the two cohorts. Area under the receiver operating characteristic curve (AUC) values for these methods ranged from 0.50 to 0.68; only three methods performed statistically better than random guessing. The radiologists’ AUC values ranged from 0.70 to 0.85; three radiologists performed statistically better than the best-performing computer method. The LUNGx Challenge compared the performance of computerized methods in the task of differentiating benign from malignant lung nodules on CT scans, placed in the context of the performance of radiologists on the same task. Lastly, the continued public availability of the Challenge cases will provide a valuable resource for the medical imaging research community.« less

A Machine Learning Ensemble Classifier for Early Prediction of Diabetic Retinopathy.

PubMed

S K, Somasundaram; P, Alli

2017-11-09

The main complication of diabetes is Diabetic retinopathy (DR), retinal vascular disease and it leads to the blindness. Regular screening for early DR disease detection is considered as an intensive labor and resource oriented task. Therefore, automatic detection of DR diseases is performed only by using the computational technique is the great solution. An automatic method is more reliable to determine the presence of an abnormality in Fundus images (FI) but, the classification process is poorly performed. Recently, few research works have been designed for analyzing texture discrimination capacity in FI to distinguish the healthy images. However, the feature extraction (FE) process was not performed well, due to the high dimensionality. Therefore, to identify retinal features for DR disease diagnosis and early detection using Machine Learning and Ensemble Classification method, called, Machine Learning Bagging Ensemble Classifier (ML-BEC) is designed. The ML-BEC method comprises of two stages. The first stage in ML-BEC method comprises extraction of the candidate objects from Retinal Images (RI). The candidate objects or the features for DR disease diagnosis include blood vessels, optic nerve, neural tissue, neuroretinal rim, optic disc size, thickness and variance. These features are initially extracted by applying Machine Learning technique called, t-distributed Stochastic Neighbor Embedding (t-SNE). Besides, t-SNE generates a probability distribution across high-dimensional images where the images are separated into similar and dissimilar pairs. Then, t-SNE describes a similar probability distribution across the points in the low-dimensional map. This lessens the Kullback-Leibler divergence among two distributions regarding the locations of the points on the map. The second stage comprises of application of ensemble classifiers to the extracted features for providing accurate analysis of digital FI using machine learning. In this stage, an automatic detection of DR screening system using Bagging Ensemble Classifier (BEC) is investigated. With the help of voting the process in ML-BEC, bagging minimizes the error due to variance of the base classifier. With the publicly available retinal image databases, our classifier is trained with 25% of RI. Results show that the ensemble classifier can achieve better classification accuracy (CA) than single classification models. Empirical experiments suggest that the machine learning-based ensemble classifier is efficient for further reducing DR classification time (CT).

Single-trial classification of motor imagery differing in task complexity: a functional near-infrared spectroscopy study

PubMed Central

2011-01-01

Background For brain computer interfaces (BCIs), which may be valuable in neurorehabilitation, brain signals derived from mental activation can be monitored by non-invasive methods, such as functional near-infrared spectroscopy (fNIRS). Single-trial classification is important for this purpose and this was the aim of the presented study. In particular, we aimed to investigate a combined approach: 1) offline single-trial classification of brain signals derived from a novel wireless fNIRS instrument; 2) to use motor imagery (MI) as mental task thereby discriminating between MI signals in response to different tasks complexities, i.e. simple and complex MI tasks. Methods 12 subjects were asked to imagine either a simple finger-tapping task using their right thumb or a complex sequential finger-tapping task using all fingers of their right hand. fNIRS was recorded over secondary motor areas of the contralateral hemisphere. Using Fisher's linear discriminant analysis (FLDA) and cross validation, we selected for each subject a best-performing feature combination consisting of 1) one out of three channel, 2) an analysis time interval ranging from 5-15 s after stimulation onset and 3) up to four Δ[O2Hb] signal features (Δ[O2Hb] mean signal amplitudes, variance, skewness and kurtosis). Results The results of our single-trial classification showed that using the simple combination set of channels, time intervals and up to four Δ[O2Hb] signal features comprising Δ[O2Hb] mean signal amplitudes, variance, skewness and kurtosis, it was possible to discriminate single-trials of MI tasks differing in complexity, i.e. simple versus complex tasks (inter-task paired t-test p ≤ 0.001), over secondary motor areas with an average classification accuracy of 81%. Conclusions Although the classification accuracies look promising they are nevertheless subject of considerable subject-to-subject variability. In the discussion we address each of these aspects, their limitations for future approaches in single-trial classification and their relevance for neurorehabilitation. PMID:21682906

A theoretical and experimental investigation of graph theoretical measures for land development in satellite imagery.

PubMed

Unsalan, Cem; Boyer, Kim L

2005-04-01

Today's commercial satellite images enable experts to classify region types in great detail. In previous work, we considered discriminating rural and urban regions [23]. However, a more detailed classification is required for many purposes. These fine classifications assist government agencies in many ways including urban planning, transportation management, and rescue operations. In a step toward the automation of the fine classification process, this paper explores graph theoretical measures over grayscale images. The graphs are constructed by assigning photometric straight line segments to vertices, while graph edges encode their spatial relationships. We then introduce a set of measures based on various properties of the graph. These measures are nearly monotonic (positively correlated) with increasing structure (organization) in the image. Thus, increased cultural activity and land development are indicated by increases in these measures-without explicit extraction of road networks, buildings, residences, etc. These latter, time consuming (and still only partially automated) tasks can be restricted only to "promising" image regions, according to our measures. In some applications our measures may suffice. We present a theoretical basis for the measures followed by extensive experimental results in which the measures are first compared to manual evaluations of land development. We then present and test a method to focus on, and (pre)extract, suburban-style residential areas. These are of particular importance in many applications, and are especially difficult to extract. In this work, we consider commercial IKONOS data. These images are orthorectified to provide a fixed resolution of 1 meter per pixel on the ground. They are, therefore, metric in the sense that ground distance is fixed in scale to pixel distance. Our data set is large and diverse, including sea and coastline, rural, forest, residential, industrial, and urban areas.

Graph Theory-Based Brain Connectivity for Automatic Classification of Multiple Sclerosis Clinical Courses.

PubMed

Kocevar, Gabriel; Stamile, Claudio; Hannoun, Salem; Cotton, François; Vukusic, Sandra; Durand-Dubief, Françoise; Sappey-Marinier, Dominique

2016-01-01

Purpose: In this work, we introduce a method to classify Multiple Sclerosis (MS) patients into four clinical profiles using structural connectivity information. For the first time, we try to solve this question in a fully automated way using a computer-based method. The main goal is to show how the combination of graph-derived metrics with machine learning techniques constitutes a powerful tool for a better characterization and classification of MS clinical profiles. Materials and Methods: Sixty-four MS patients [12 Clinical Isolated Syndrome (CIS), 24 Relapsing Remitting (RR), 24 Secondary Progressive (SP), and 17 Primary Progressive (PP)] along with 26 healthy controls (HC) underwent MR examination. T1 and diffusion tensor imaging (DTI) were used to obtain structural connectivity matrices for each subject. Global graph metrics, such as density and modularity, were estimated and compared between subjects' groups. These metrics were further used to classify patients using tuned Support Vector Machine (SVM) combined with Radial Basic Function (RBF) kernel. Results: When comparing MS patients to HC subjects, a greater assortativity, transitivity, and characteristic path length as well as a lower global efficiency were found. Using all graph metrics, the best F -Measures (91.8, 91.8, 75.6, and 70.6%) were obtained for binary (HC-CIS, CIS-RR, RR-PP) and multi-class (CIS-RR-SP) classification tasks, respectively. When using only one graph metric, the best F -Measures (83.6, 88.9, and 70.7%) were achieved for modularity with previous binary classification tasks. Conclusion: Based on a simple DTI acquisition associated with structural brain connectivity analysis, this automatic method allowed an accurate classification of different MS patients' clinical profiles.

A Deep Convolutional Neural Network for segmenting and classifying epithelial and stromal regions in histopathological images

PubMed Central

Xu, Jun; Luo, Xiaofei; Wang, Guanhao; Gilmore, Hannah; Madabhushi, Anant

2016-01-01

Epithelial (EP) and stromal (ST) are two types of tissues in histological images. Automated segmentation or classification of EP and ST tissues is important when developing computerized system for analyzing the tumor microenvironment. In this paper, a Deep Convolutional Neural Networks (DCNN) based feature learning is presented to automatically segment or classify EP and ST regions from digitized tumor tissue microarrays (TMAs). Current approaches are based on handcraft feature representation, such as color, texture, and Local Binary Patterns (LBP) in classifying two regions. Compared to handcrafted feature based approaches, which involve task dependent representation, DCNN is an end-to-end feature extractor that may be directly learned from the raw pixel intensity value of EP and ST tissues in a data driven fashion. These high-level features contribute to the construction of a supervised classifier for discriminating the two types of tissues. In this work we compare DCNN based models with three handcraft feature extraction based approaches on two different datasets which consist of 157 Hematoxylin and Eosin (H&E) stained images of breast cancer and 1376 immunohistological (IHC) stained images of colorectal cancer, respectively. The DCNN based feature learning approach was shown to have a F1 classification score of 85%, 89%, and 100%, accuracy (ACC) of 84%, 88%, and 100%, and Matthews Correlation Coefficient (MCC) of 86%, 77%, and 100% on two H&E stained (NKI and VGH) and IHC stained data, respectively. Our DNN based approach was shown to outperform three handcraft feature extraction based approaches in terms of the classification of EP and ST regions. PMID:28154470

A Deep Convolutional Neural Network for segmenting and classifying epithelial and stromal regions in histopathological images.

PubMed

Xu, Jun; Luo, Xiaofei; Wang, Guanhao; Gilmore, Hannah; Madabhushi, Anant

2016-05-26

Epithelial (EP) and stromal (ST) are two types of tissues in histological images. Automated segmentation or classification of EP and ST tissues is important when developing computerized system for analyzing the tumor microenvironment. In this paper, a Deep Convolutional Neural Networks (DCNN) based feature learning is presented to automatically segment or classify EP and ST regions from digitized tumor tissue microarrays (TMAs). Current approaches are based on handcraft feature representation, such as color, texture, and Local Binary Patterns (LBP) in classifying two regions. Compared to handcrafted feature based approaches, which involve task dependent representation, DCNN is an end-to-end feature extractor that may be directly learned from the raw pixel intensity value of EP and ST tissues in a data driven fashion. These high-level features contribute to the construction of a supervised classifier for discriminating the two types of tissues. In this work we compare DCNN based models with three handcraft feature extraction based approaches on two different datasets which consist of 157 Hematoxylin and Eosin (H&E) stained images of breast cancer and 1376 immunohistological (IHC) stained images of colorectal cancer, respectively. The DCNN based feature learning approach was shown to have a F1 classification score of 85%, 89%, and 100%, accuracy (ACC) of 84%, 88%, and 100%, and Matthews Correlation Coefficient (MCC) of 86%, 77%, and 100% on two H&E stained (NKI and VGH) and IHC stained data, respectively. Our DNN based approach was shown to outperform three handcraft feature extraction based approaches in terms of the classification of EP and ST regions.

Haptic fMRI: using classification to quantify task-correlated noise during goal-directed reaching motions.

PubMed

Menon, Samir; Quigley, Paul; Yu, Michelle; Khatib, Oussama

2014-01-01

Neuroimaging artifacts in haptic functional magnetic resonance imaging (Haptic fMRI) experiments have the potential to induce spurious fMRI activation where there is none, or to make neural activation measurements appear correlated across brain regions when they are actually not. Here, we demonstrate that performing three-dimensional goal-directed reaching motions while operating Haptic fMRI Interface (HFI) does not create confounding motion artifacts. To test for artifacts, we simultaneously scanned a subject's brain with a customized soft phantom placed a few centimeters away from the subject's left motor cortex. The phantom captured task-related motion and haptic noise, but did not contain associated neural activation measurements. We quantified the task-related information present in fMRI measurements taken from the brain and the phantom by using a linear max-margin classifier to predict whether raw time series data could differentiate between motion planning or reaching. fMRI measurements in the phantom were uninformative (2σ, 45-73%; chance=50%), while those in primary motor, visual, and somatosensory cortex accurately classified task-conditions (2σ, 90-96%). We also localized artifacts due to the haptic interface alone by scanning a stand-alone fBIRN phantom, while an operator performed haptic tasks outside the scanner's bore with the interface at the same location. The stand-alone phantom had lower temporal noise and had similar mean classification but a tighter distribution (bootstrap Gaussian fit) than the brain phantom. Our results suggest that any fMRI measurement artifacts for Haptic fMRI reaching experiments are dominated by actual neural responses.

Object based technique for delineating and mapping 15 tree species using VHR WorldView-2 imagery

NASA Astrophysics Data System (ADS)

Mustafa, Yaseen T.; Habeeb, Hindav N.

2014-10-01

Monitoring and analyzing forests and trees are required task to manage and establish a good plan for the forest sustainability. To achieve such a task, information and data collection of the trees are requested. The fastest way and relatively low cost technique is by using satellite remote sensing. In this study, we proposed an approach to identify and map 15 tree species in the Mangish sub-district, Kurdistan Region-Iraq. Image-objects (IOs) were used as the tree species mapping unit. This is achieved using the shadow index, normalized difference vegetation index and texture measurements. Four classification methods (Maximum Likelihood, Mahalanobis Distance, Neural Network, and Spectral Angel Mapper) were used to classify IOs using selected IO features derived from WorldView-2 imagery. Results showed that overall accuracy was increased 5-8% using the Neural Network method compared with other methods with a Kappa coefficient of 69%. This technique gives reasonable results of various tree species classifications by means of applying the Neural Network method with IOs techniques on WorldView-2 imagery.

A latent discriminative model-based approach for classification of imaginary motor tasks from EEG data.

PubMed

Saa, Jaime F Delgado; Çetin, Müjdat

2012-04-01

We consider the problem of classification of imaginary motor tasks from electroencephalography (EEG) data for brain-computer interfaces (BCIs) and propose a new approach based on hidden conditional random fields (HCRFs). HCRFs are discriminative graphical models that are attractive for this problem because they (1) exploit the temporal structure of EEG; (2) include latent variables that can be used to model different brain states in the signal; and (3) involve learned statistical models matched to the classification task, avoiding some of the limitations of generative models. Our approach involves spatial filtering of the EEG signals and estimation of power spectra based on autoregressive modeling of temporal segments of the EEG signals. Given this time-frequency representation, we select certain frequency bands that are known to be associated with execution of motor tasks. These selected features constitute the data that are fed to the HCRF, parameters of which are learned from training data. Inference algorithms on the HCRFs are used for the classification of motor tasks. We experimentally compare this approach to the best performing methods in BCI competition IV as well as a number of more recent methods and observe that our proposed method yields better classification accuracy.

Mental Task Classification Scheme Utilizing Correlation Coefficient Extracted from Interchannel Intrinsic Mode Function.

PubMed

Rahman, Md Mostafizur; Fattah, Shaikh Anowarul

2017-01-01

In view of recent increase of brain computer interface (BCI) based applications, the importance of efficient classification of various mental tasks has increased prodigiously nowadays. In order to obtain effective classification, efficient feature extraction scheme is necessary, for which, in the proposed method, the interchannel relationship among electroencephalogram (EEG) data is utilized. It is expected that the correlation obtained from different combination of channels will be different for different mental tasks, which can be exploited to extract distinctive feature. The empirical mode decomposition (EMD) technique is employed on a test EEG signal obtained from a channel, which provides a number of intrinsic mode functions (IMFs), and correlation coefficient is extracted from interchannel IMF data. Simultaneously, different statistical features are also obtained from each IMF. Finally, the feature matrix is formed utilizing interchannel correlation features and intrachannel statistical features of the selected IMFs of EEG signal. Different kernels of the support vector machine (SVM) classifier are used to carry out the classification task. An EEG dataset containing ten different combinations of five different mental tasks is utilized to demonstrate the classification performance and a very high level of accuracy is achieved by the proposed scheme compared to existing methods.

A New Pivoting and Iterative Text Detection Algorithm for Biomedical Images

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

Xu, Songhua; Krauthammer, Prof. Michael

2010-01-01

There is interest to expand the reach of literature mining to include the analysis of biomedical images, which often contain a paper's key findings. Examples include recent studies that use Optical Character Recognition (OCR) to extract image text, which is used to boost biomedical image retrieval and classification. Such studies rely on the robust identification of text elements in biomedical images, which is a non-trivial task. In this work, we introduce a new text detection algorithm for biomedical images based on iterative projection histograms. We study the effectiveness of our algorithm by evaluating the performance on a set of manuallymore » labeled random biomedical images, and compare the performance against other state-of-the-art text detection algorithms. We demonstrate that our projection histogram-based text detection approach is well suited for text detection in biomedical images, and that the iterative application of the algorithm boosts performance to an F score of .60. We provide a C++ implementation of our algorithm freely available for academic use.« less

An Active Patch Model for Real World Texture and Appearance Classification

PubMed Central

Mao, Junhua; Zhu, Jun; Yuille, Alan L.

2014-01-01

This paper addresses the task of natural texture and appearance classification. Our goal is to develop a simple and intuitive method that performs at state of the art on datasets ranging from homogeneous texture (e.g., material texture), to less homogeneous texture (e.g., the fur of animals), and to inhomogeneous texture (the appearance patterns of vehicles). Our method uses a bag-of-words model where the features are based on a dictionary of active patches. Active patches are raw intensity patches which can undergo spatial transformations (e.g., rotation and scaling) and adjust themselves to best match the image regions. The dictionary of active patches is required to be compact and representative, in the sense that we can use it to approximately reconstruct the images that we want to classify. We propose a probabilistic model to quantify the quality of image reconstruction and design a greedy learning algorithm to obtain the dictionary. We classify images using the occurrence frequency of the active patches. Feature extraction is fast (about 100 ms per image) using the GPU. The experimental results show that our method improves the state of the art on a challenging material texture benchmark dataset (KTH-TIPS2). To test our method on less homogeneous or inhomogeneous images, we construct two new datasets consisting of appearance image patches of animals and vehicles cropped from the PASCAL VOC dataset. Our method outperforms competing methods on these datasets. PMID:25531013

Intelligent MRTD testing for thermal imaging system using ANN

NASA Astrophysics Data System (ADS)

Sun, Junyue; Ma, Dongmei

2006-01-01

The Minimum Resolvable Temperature Difference (MRTD) is the most widely accepted figure for describing the performance of a thermal imaging system. Many models have been proposed to predict it. The MRTD testing is a psychophysical task, for which biases are unavoidable. It requires laboratory conditions such as normal air condition and a constant temperature. It also needs expensive measuring equipments and takes a considerable period of time. Especially when measuring imagers of the same type, the test is time consuming. So an automated and intelligent measurement method should be discussed. This paper adopts the concept of automated MRTD testing using boundary contour system and fuzzy ARTMAP, but uses different methods. It describes an Automated MRTD Testing procedure basing on Back-Propagation Network. Firstly, we use frame grabber to capture the 4-bar target image data. Then according to image gray scale, we segment the image to get 4-bar place and extract feature vector representing the image characteristic and human detection ability. These feature sets, along with known target visibility, are used to train the ANN (Artificial Neural Networks). Actually it is a nonlinear classification (of input dimensions) of the image series using ANN. Our task is to justify if image is resolvable or uncertainty. Then the trained ANN will emulate observer performance in determining MRTD. This method can reduce the uncertainties between observers and long time dependent factors by standardization. This paper will introduce the feature extraction algorithm, demonstrate the feasibility of the whole process and give the accuracy of MRTD measurement.

Analysis of the hand vein pattern for people recognition

NASA Astrophysics Data System (ADS)

Castro-Ortega, R.; Toxqui-Quitl, C.; Cristóbal, G.; Marcos, J. Victor; Padilla-Vivanco, A.; Hurtado Pérez, R.

2015-09-01

The shape of the hand vascular pattern contains useful and unique features that can be used for identifying and authenticating people, with applications in access control, medicine and financial services. In this work, an optical system for the image acquisition of the hand vascular pattern is implemented. It consists of a CCD camera with sensitivity in the IR and a light source with emission in the 880 nm. The IR radiation interacts with the desoxyhemoglobin, hemoglobin and water present in the blood of the veins, making possible to see the vein pattern underneath skin. The segmentation of the Region Of Interest (ROI) is achieved using geometrical moments locating the centroid of an image. For enhancement of the vein pattern we use the technique of Histogram Equalization and Contrast Limited Adaptive Histogram Equalization (CLAHE). In order to remove unnecessary information such as body hair and skinfolds, a low pass filter is implemented. A method based on geometric moments is used to obtain the invariant descriptors of the input images. The classification task is achieved using Artificial Neural Networks (ANN) and K-Nearest Neighbors (K-nn) algorithms. Experimental results using our database show a percentage of correct classification, higher of 86.36% with ANN for 912 images of 38 people with 12 versions each one.

Classification of micro-CT images using 3D characterization of bone canal patterns in human osteogenesis imperfecta

NASA Astrophysics Data System (ADS)

Abidin, Anas Z.; Jameson, John; Molthen, Robert; Wismüller, Axel

2017-03-01

Few studies have analyzed the microstructural properties of bone in cases of Osteogenenis Imperfecta (OI), or `brittle bone disease'. Current approaches mainly focus on bone mineral density measurements as an indirect indicator of bone strength and quality. It has been shown that bone strength would depend not only on composition but also structural organization. This study aims to characterize 3D structure of the cortical bone in high-resolution micro CT images. A total of 40 bone fragments from 28 subjects (13 with OI and 15 healthy controls) were imaged using micro tomography using a synchrotron light source (SRµCT). Minkowski functionals - volume, surface, curvature, and Euler characteristics - describing the topological organization of the bone were computed from the images. The features were used in a machine learning task to classify between healthy and OI bone. The best classification performance (mean AUC - 0.96) was achieved with a combined 4-dimensional feature of all Minkowski functionals. Individually, the best feature performance was seen using curvature (mean AUC - 0.85), which characterizes the edges within a binary object. These results show that quantitative analysis of cortical bone microstructure, in a computer-aided diagnostics framework, can be used to distinguish between healthy and OI bone with high accuracy.

An incremental knowledge assimilation system (IKAS) for mine detection

NASA Astrophysics Data System (ADS)

Porway, Jake; Raju, Chaitanya; Varadarajan, Karthik Mahesh; Nguyen, Hieu; Yadegar, Joseph

2010-04-01

In this paper we present an adaptive incremental learning system for underwater mine detection and classification that utilizes statistical models of seabed texture and an adaptive nearest-neighbor classifier to identify varied underwater targets in many different environments. The first stage of processing uses our Background Adaptive ANomaly detector (BAAN), which identifies statistically likely target regions using Gabor filter responses over the image. Using this information, BAAN classifies the background type and updates its detection using background-specific parameters. To perform classification, a Fully Adaptive Nearest Neighbor (FAAN) determines the best label for each detection. FAAN uses an extremely fast version of Nearest Neighbor to find the most likely label for the target. The classifier perpetually assimilates new and relevant information into its existing knowledge database in an incremental fashion, allowing improved classification accuracy and capturing concept drift in the target classes. Experiments show that the system achieves >90% classification accuracy on underwater mine detection tasks performed on synthesized datasets provided by the Office of Naval Research. We have also demonstrated that the system can incrementally improve its detection accuracy by constantly learning from new samples.

Perceptual-motor skill learning in Gilles de la Tourette syndrome. Evidence for multiple procedural learning and memory systems.

PubMed

Marsh, Rachel; Alexander, Gerianne M; Packard, Mark G; Zhu, Hongtu; Peterson, Bradley S

2005-01-01

Procedural learning and memory systems likely comprise several skills that are differentially affected by various illnesses of the central nervous system, suggesting their relative functional independence and reliance on differing neural circuits. Gilles de la Tourette syndrome (GTS) is a movement disorder that involves disturbances in the structure and function of the striatum and related circuitry. Recent studies suggest that patients with GTS are impaired in performance of a probabilistic classification task that putatively involves the acquisition of stimulus-response (S-R)-based habits. Assessing the learning of perceptual-motor skills and probabilistic classification in the same samples of GTS and healthy control subjects may help to determine whether these various forms of procedural (habit) learning rely on the same or differing neuroanatomical substrates and whether those substrates are differentially affected in persons with GTS. Therefore, we assessed perceptual-motor skill learning using the pursuit-rotor and mirror tracing tasks in 50 patients with GTS and 55 control subjects who had previously been compared at learning a task of probabilistic classifications. The GTS subjects did not differ from the control subjects in performance of either the pursuit rotor or mirror-tracing tasks, although they were significantly impaired in the acquisition of a probabilistic classification task. In addition, learning on the perceptual-motor tasks was not correlated with habit learning on the classification task in either the GTS or healthy control subjects. These findings suggest that the differing forms of procedural learning are dissociable both functionally and neuroanatomically. The specific deficits in the probabilistic classification form of habit learning in persons with GTS are likely to be a consequence of disturbances in specific corticostriatal circuits, but not the same circuits that subserve the perceptual-motor form of habit learning.

Is overall similarity classification less effortful than single-dimension classification?

PubMed

Wills, Andy J; Milton, Fraser; Longmore, Christopher A; Hester, Sarah; Robinson, Jo

2013-01-01

It is sometimes argued that the implementation of an overall similarity classification is less effortful than the implementation of a single-dimension classification. In the current article, we argue that the evidence securely in support of this view is limited, and report additional evidence in support of the opposite proposition--overall similarity classification is more effortful than single-dimension classification. Using a match-to-standards procedure, Experiments 1A, 1B and 2 demonstrate that concurrent load reduces the prevalence of overall similarity classification, and that this effect is robust to changes in the concurrent load task employed, the level of time pressure experienced, and the short-term memory requirements of the classification task. Experiment 3 demonstrates that participants who produced overall similarity classifications from the outset have larger working memory capacities than those who produced single-dimension classifications initially, and Experiment 4 demonstrates that instructions to respond meticulously increase the prevalence of overall similarity classification.

Psoriasis skin biopsy image segmentation using Deep Convolutional Neural Network.

PubMed

Pal, Anabik; Garain, Utpal; Chandra, Aditi; Chatterjee, Raghunath; Senapati, Swapan

2018-06-01

Development of machine assisted tools for automatic analysis of psoriasis skin biopsy image plays an important role in clinical assistance. Development of automatic approach for accurate segmentation of psoriasis skin biopsy image is the initial prerequisite for developing such system. However, the complex cellular structure, presence of imaging artifacts, uneven staining variation make the task challenging. This paper presents a pioneering attempt for automatic segmentation of psoriasis skin biopsy images. Several deep neural architectures are tried for segmenting psoriasis skin biopsy images. Deep models are used for classifying the super-pixels generated by Simple Linear Iterative Clustering (SLIC) and the segmentation performance of these architectures is compared with the traditional hand-crafted feature based classifiers built on popularly used classifiers like K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and Random Forest (RF). A U-shaped Fully Convolutional Neural Network (FCN) is also used in an end to end learning fashion where input is the original color image and the output is the segmentation class map for the skin layers. An annotated real psoriasis skin biopsy image data set of ninety (90) images is developed and used for this research. The segmentation performance is evaluated with two metrics namely, Jaccard's Coefficient (JC) and the Ratio of Correct Pixel Classification (RCPC) accuracy. The experimental results show that the CNN based approaches outperform the traditional hand-crafted feature based classification approaches. The present research shows that practical system can be developed for machine assisted analysis of psoriasis disease. Copyright © 2018 Elsevier B.V. All rights reserved.

Integrating Dimension Reduction and Out-of-Sample Extension in Automated Classification of Ex Vivo Human Patellar Cartilage on Phase Contrast X-Ray Computed Tomography

PubMed Central

Nagarajan, Mahesh B.; Coan, Paola; Huber, Markus B.; Diemoz, Paul C.; Wismüller, Axel

2015-01-01

Phase contrast X-ray computed tomography (PCI-CT) has been demonstrated as a novel imaging technique that can visualize human cartilage with high spatial resolution and soft tissue contrast. Different textural approaches have been previously investigated for characterizing chondrocyte organization on PCI-CT to enable classification of healthy and osteoarthritic cartilage. However, the large size of feature sets extracted in such studies motivates an investigation into algorithmic feature reduction for computing efficient feature representations without compromising their discriminatory power. For this purpose, geometrical feature sets derived from the scaling index method (SIM) were extracted from 1392 volumes of interest (VOI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. The extracted feature sets were subject to linear and non-linear dimension reduction techniques as well as feature selection based on evaluation of mutual information criteria. The reduced feature set was subsequently used in a machine learning task with support vector regression to classify VOIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver-operating characteristic (ROC) curve (AUC). Our results show that the classification performance achieved by 9-D SIM-derived geometric feature sets (AUC: 0.96 ± 0.02) can be maintained with 2-D representations computed from both dimension reduction and feature selection (AUC values as high as 0.97 ± 0.02). Thus, such feature reduction techniques can offer a high degree of compaction to large feature sets extracted from PCI-CT images while maintaining their ability to characterize the underlying chondrocyte patterns. PMID:25710875

Stratification of pseudoprogression and true progression of glioblastoma multiform based on longitudinal diffusion tensor imaging without segmentation

PubMed Central

Qian, Xiaohua; Tan, Hua; Zhang, Jian; Zhao, Weilin; Chan, Michael D.; Zhou, Xiaobo

2016-01-01

Purpose: Pseudoprogression (PsP) can mimic true tumor progression (TTP) on magnetic resonance imaging in patients with glioblastoma multiform (GBM). The phenotypical similarity between PsP and TTP makes it a challenging task for physicians to distinguish these entities. So far, no approved biomarkers or computer-aided diagnosis systems have been used clinically for this purpose. Methods: To address this challenge, the authors developed an objective classification system for PsP and TTP based on longitudinal diffusion tensor imaging. A novel spatio-temporal discriminative dictionary learning scheme was proposed to differentiate PsP and TTP, thereby avoiding segmentation of the region of interest. The authors constructed a novel discriminative sparse matrix with the classification-oriented dictionary learning approach by excluding the shared features of two categories, so that the pooled features captured the subtle difference between PsP and TTP. The most discriminating features were then identified from the pooled features by their feature scoring system. Finally, the authors stratified patients with GBM into PsP and TTP by a support vector machine approach. Tenfold cross-validation (CV) and the area under the receiver operating characteristic (AUC) were used to assess the robustness of the developed system. Results: The average accuracy and AUC values after ten rounds of tenfold CV were 0.867 and 0.92, respectively. The authors also assessed the effects of different methods and factors (such as data types, pooling techniques, and dimensionality reduction approaches) on the performance of their classification system which obtained the best performance. Conclusions: The proposed objective classification system without segmentation achieved a desirable and reliable performance in differentiating PsP from TTP. Thus, the developed approach is expected to advance the clinical research and diagnosis of PsP and TTP. PMID:27806598

Treelets Binary Feature Retrieval for Fast Keypoint Recognition.

PubMed

Zhu, Jianke; Wu, Chenxia; Chen, Chun; Cai, Deng

2015-10-01

Fast keypoint recognition is essential to many vision tasks. In contrast to the classification-based approaches, we directly formulate the keypoint recognition as an image patch retrieval problem, which enjoys the merit of finding the matched keypoint and its pose simultaneously. To effectively extract the binary features from each patch surrounding the keypoint, we make use of treelets transform that can group the highly correlated data together and reduce the noise through the local analysis. Treelets is a multiresolution analysis tool, which provides an orthogonal basis to reflect the geometry of the noise-free data. To facilitate the real-world applications, we have proposed two novel approaches. One is the convolutional treelets that capture the image patch information locally and globally while reducing the computational cost. The other is the higher-order treelets that reflect the relationship between the rows and columns within image patch. An efficient sub-signature-based locality sensitive hashing scheme is employed for fast approximate nearest neighbor search in patch retrieval. Experimental evaluations on both synthetic data and the real-world Oxford dataset have shown that our proposed treelets binary feature retrieval methods outperform the state-of-the-art feature descriptors and classification-based approaches.

An Automatic Diagnosis Method of Facial Acne Vulgaris Based on Convolutional Neural Network.

PubMed

Shen, Xiaolei; Zhang, Jiachi; Yan, Chenjun; Zhou, Hong

2018-04-11

In this paper, we present a new automatic diagnosis method for facial acne vulgaris which is based on convolutional neural networks (CNNs). To overcome the shortcomings of previous methods which were the inability to classify enough types of acne vulgaris. The core of our method is to extract features of images based on CNNs and achieve classification by classifier. A binary-classifier of skin-and-non-skin is used to detect skin area and a seven-classifier is used to achieve the classification task of facial acne vulgaris and healthy skin. In the experiments, we compare the effectiveness of our CNN and the VGG16 neural network which is pre-trained on the ImageNet data set. We use a ROC curve to evaluate the performance of binary-classifier and use a normalized confusion matrix to evaluate the performance of seven-classifier. The results of our experiments show that the pre-trained VGG16 neural network is effective in extracting features from facial acne vulgaris images. And the features are very useful for the follow-up classifiers. Finally, we try applying the classifiers both based on the pre-trained VGG16 neural network to assist doctors in facial acne vulgaris diagnosis.

Breast cancer detection in rotational thermography images using texture features

NASA Astrophysics Data System (ADS)

Francis, Sheeja V.; Sasikala, M.; Bhavani Bharathi, G.; Jaipurkar, Sandeep D.

2014-11-01

Breast cancer is a major cause of mortality in young women in the developing countries. Early diagnosis is the key to improve survival rate in cancer patients. Breast thermography is a diagnostic procedure that non-invasively images the infrared emissions from breast surface to aid in the early detection of breast cancer. Due to limitations in imaging protocol, abnormality detection by conventional breast thermography, is often a challenging task. Rotational thermography is a novel technique developed in order to overcome the limitations of conventional breast thermography. This paper evaluates this technique's potential for automatic detection of breast abnormality, from the perspective of cold challenge. Texture features are extracted in the spatial domain, from rotational thermogram series, prior to and post the application of cold challenge. These features are fed to a support vector machine for automatic classification of normal and malignant breasts, resulting in a classification accuracy of 83.3%. Feature reduction has been performed by principal component analysis. As a novel attempt, the ability of this technique to locate the abnormality has been studied. The results of the study indicate that rotational thermography holds great potential as a screening tool for breast cancer detection.

A Framework for Architecture-Based Planning and Assessment to Support Modeling and Simulation of Network-Centric Command and Control

DTIC Science & Technology

2006-02-17

included system-unique specifications derived to counter these threats. But the international security environment has changed --- and it will continue... Netowrk -Centric Command and Control 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK...contains color images. 14. ABSTRACT 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT 18. NUMBER OF PAGES 20 19a. NAME

Objected-oriented remote sensing image classification method based on geographic ontology model

NASA Astrophysics Data System (ADS)

Chu, Z.; Liu, Z. J.; Gu, H. Y.

2016-11-01

Nowadays, with the development of high resolution remote sensing image and the wide application of laser point cloud data, proceeding objected-oriented remote sensing classification based on the characteristic knowledge of multi-source spatial data has been an important trend on the field of remote sensing image classification, which gradually replaced the traditional method through improving algorithm to optimize image classification results. For this purpose, the paper puts forward a remote sensing image classification method that uses the he characteristic knowledge of multi-source spatial data to build the geographic ontology semantic network model, and carries out the objected-oriented classification experiment to implement urban features classification, the experiment uses protégé software which is developed by Stanford University in the United States, and intelligent image analysis software—eCognition software as the experiment platform, uses hyperspectral image and Lidar data that is obtained through flight in DaFeng City of JiangSu as the main data source, first of all, the experiment uses hyperspectral image to obtain feature knowledge of remote sensing image and related special index, the second, the experiment uses Lidar data to generate nDSM(Normalized DSM, Normalized Digital Surface Model),obtaining elevation information, the last, the experiment bases image feature knowledge, special index and elevation information to build the geographic ontology semantic network model that implement urban features classification, the experiment results show that, this method is significantly higher than the traditional classification algorithm on classification accuracy, especially it performs more evidently on the respect of building classification. The method not only considers the advantage of multi-source spatial data, for example, remote sensing image, Lidar data and so on, but also realizes multi-source spatial data knowledge integration and application of the knowledge to the field of remote sensing image classification, which provides an effective way for objected-oriented remote sensing image classification in the future.

Classification of Self-Driven Mental Tasks from Whole-Brain Activity Patterns

PubMed Central

Nawa, Norberto Eiji; Ando, Hiroshi

2014-01-01

During wakefulness, a constant and continuous stream of complex stimuli and self-driven thoughts permeate the human mind. Here, eleven participants were asked to count down numbers and remember negative or positive autobiographical episodes of their personal lives, for 32 seconds at a time, during which they could freely engage in the execution of those tasks. We then examined the possibility of determining from a single whole-brain functional magnetic resonance imaging scan which one of the two mental tasks each participant was performing at a given point in time. Linear support-vector machines were used to build within-participant classifiers and across-participants classifiers. The within-participant classifiers could correctly discriminate scans with an average accuracy as high as 82%, when using data from all individual voxels in the brain. These results demonstrate that it is possible to accurately classify self-driven mental tasks from whole-brain activity patterns recorded in a time interval as short as 2 seconds. PMID:24824899

Generating Text from Functional Brain Images

PubMed Central

Pereira, Francisco; Detre, Greg; Botvinick, Matthew

2011-01-01

Recent work has shown that it is possible to take brain images acquired during viewing of a scene and reconstruct an approximation of the scene from those images. Here we show that it is also possible to generate text about the mental content reflected in brain images. We began with images collected as participants read names of concrete items (e.g., “Apartment’’) while also seeing line drawings of the item named. We built a model of the mental semantic representation of concrete concepts from text data and learned to map aspects of such representation to patterns of activation in the corresponding brain image. In order to validate this mapping, without accessing information about the items viewed for left-out individual brain images, we were able to generate from each one a collection of semantically pertinent words (e.g., “door,” “window” for “Apartment’’). Furthermore, we show that the ability to generate such words allows us to perform a classification task and thus validate our method quantitatively. PMID:21927602

Computer-aided diagnosis for phase-contrast X-ray computed tomography: quantitative characterization of human patellar cartilage with high-dimensional geometric features.

PubMed

Nagarajan, Mahesh B; Coan, Paola; Huber, Markus B; Diemoz, Paul C; Glaser, Christian; Wismüller, Axel

2014-02-01

Phase-contrast computed tomography (PCI-CT) has shown tremendous potential as an imaging modality for visualizing human cartilage with high spatial resolution. Previous studies have demonstrated the ability of PCI-CT to visualize (1) structural details of the human patellar cartilage matrix and (2) changes to chondrocyte organization induced by osteoarthritis. This study investigates the use of high-dimensional geometric features in characterizing such chondrocyte patterns in the presence or absence of osteoarthritic damage. Geometrical features derived from the scaling index method (SIM) and statistical features derived from gray-level co-occurrence matrices were extracted from 842 regions of interest (ROI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. These features were subsequently used in a machine learning task with support vector regression to classify ROIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver-operating characteristic curve (AUC). SIM-derived geometrical features exhibited the best classification performance (AUC, 0.95 ± 0.06) and were most robust to changes in ROI size. These results suggest that such geometrical features can provide a detailed characterization of the chondrocyte organization in the cartilage matrix in an automated and non-subjective manner, while also enabling classification of cartilage as healthy or osteoarthritic with high accuracy. Such features could potentially serve as imaging markers for evaluating osteoarthritis progression and its response to different therapeutic intervention strategies.

Task-specific feature extraction and classification of fMRI volumes using a deep neural network initialized with a deep belief network: Evaluation using sensorimotor tasks

PubMed Central

Jang, Hojin; Plis, Sergey M.; Calhoun, Vince D.; Lee, Jong-Hwan

2016-01-01

Feedforward deep neural networks (DNN), artificial neural networks with multiple hidden layers, have recently demonstrated a record-breaking performance in multiple areas of applications in computer vision and speech processing. Following the success, DNNs have been applied to neuroimaging modalities including functional/structural magnetic resonance imaging (MRI) and positron-emission tomography data. However, no study has explicitly applied DNNs to 3D whole-brain fMRI volumes and thereby extracted hidden volumetric representations of fMRI that are discriminative for a task performed as the fMRI volume was acquired. Our study applied fully connected feedforward DNN to fMRI volumes collected in four sensorimotor tasks (i.e., left-hand clenching, right-hand clenching, auditory attention, and visual stimulus) undertaken by 12 healthy participants. Using a leave-one-subject-out cross-validation scheme, a restricted Boltzmann machine-based deep belief network was pretrained and used to initialize weights of the DNN. The pretrained DNN was fine-tuned while systematically controlling weight-sparsity levels across hidden layers. Optimal weight-sparsity levels were determined from a minimum validation error rate of fMRI volume classification. Minimum error rates (mean ± standard deviation; %) of 6.9 (± 3.8) were obtained from the three-layer DNN with the sparsest condition of weights across the three hidden layers. These error rates were even lower than the error rates from the single-layer network (9.4 ± 4.6) and the two-layer network (7.4 ± 4.1). The estimated DNN weights showed spatial patterns that are remarkably task-specific, particularly in the higher layers. The output values of the third hidden layer represented distinct patterns/codes of the 3D whole-brain fMRI volume and encoded the information of the tasks as evaluated from representational similarity analysis. Our reported findings show the ability of the DNN to classify a single fMRI volume based on the extraction of hidden representations of fMRI volumes associated with tasks across multiple hidden layers. Our study may be beneficial to the automatic classification/diagnosis of neuropsychiatric and neurological diseases and prediction of disease severity and recovery in (pre-) clinical settings using fMRI volumes without requiring an estimation of activation patterns or ad hoc statistical evaluation. PMID:27079534

Task-specific feature extraction and classification of fMRI volumes using a deep neural network initialized with a deep belief network: Evaluation using sensorimotor tasks.

PubMed

Jang, Hojin; Plis, Sergey M; Calhoun, Vince D; Lee, Jong-Hwan

2017-01-15

Feedforward deep neural networks (DNNs), artificial neural networks with multiple hidden layers, have recently demonstrated a record-breaking performance in multiple areas of applications in computer vision and speech processing. Following the success, DNNs have been applied to neuroimaging modalities including functional/structural magnetic resonance imaging (MRI) and positron-emission tomography data. However, no study has explicitly applied DNNs to 3D whole-brain fMRI volumes and thereby extracted hidden volumetric representations of fMRI that are discriminative for a task performed as the fMRI volume was acquired. Our study applied fully connected feedforward DNN to fMRI volumes collected in four sensorimotor tasks (i.e., left-hand clenching, right-hand clenching, auditory attention, and visual stimulus) undertaken by 12 healthy participants. Using a leave-one-subject-out cross-validation scheme, a restricted Boltzmann machine-based deep belief network was pretrained and used to initialize weights of the DNN. The pretrained DNN was fine-tuned while systematically controlling weight-sparsity levels across hidden layers. Optimal weight-sparsity levels were determined from a minimum validation error rate of fMRI volume classification. Minimum error rates (mean±standard deviation; %) of 6.9 (±3.8) were obtained from the three-layer DNN with the sparsest condition of weights across the three hidden layers. These error rates were even lower than the error rates from the single-layer network (9.4±4.6) and the two-layer network (7.4±4.1). The estimated DNN weights showed spatial patterns that are remarkably task-specific, particularly in the higher layers. The output values of the third hidden layer represented distinct patterns/codes of the 3D whole-brain fMRI volume and encoded the information of the tasks as evaluated from representational similarity analysis. Our reported findings show the ability of the DNN to classify a single fMRI volume based on the extraction of hidden representations of fMRI volumes associated with tasks across multiple hidden layers. Our study may be beneficial to the automatic classification/diagnosis of neuropsychiatric and neurological diseases and prediction of disease severity and recovery in (pre-) clinical settings using fMRI volumes without requiring an estimation of activation patterns or ad hoc statistical evaluation. Copyright © 2016 Elsevier Inc. All rights reserved.

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

Characterizing cartilage microarchitecture on phase-contrast x-ray computed tomography using deep learning with convolutional neural networks

NASA Astrophysics Data System (ADS)

Deng, Botao; Abidin, Anas Z.; D'Souza, Adora M.; Nagarajan, Mahesh B.; Coan, Paola; Wismüller, Axel

2017-03-01

The effectiveness of phase contrast X-ray computed tomography (PCI-CT) in visualizing human patellar cartilage matrix has been demonstrated due to its ability to capture soft tissue contrast on a micrometer resolution scale. Recent studies have shown that off-the-shelf Convolutional Neural Network (CNN) features learned from a nonmedical data set can be used for medical image classification. In this paper, we investigate the ability of features extracted from two different CNNs for characterizing chondrocyte patterns in the cartilage matrix. We obtained features from 842 regions of interest annotated on PCI-CT images of human patellar cartilage using CaffeNet and Inception-v3 Network, which were then used in a machine learning task involving support vector machines with radial basis function kernel to classify the ROIs as healthy or osteoarthritic. Classification performance was evaluated using the area (AUC) under the Receiver Operating Characteristic (ROC) curve. The best classification performance was observed with features from Inception-v3 network (AUC = 0.95), which outperforms features extracted from CaffeNet (AUC = 0.91). These results suggest that such characterization of chondrocyte patterns using features from internal layers of CNNs can be used to distinguish between healthy and osteoarthritic tissue with high accuracy.

Spectral unmixing of urban land cover using a generic library approach

NASA Astrophysics Data System (ADS)

Degerickx, Jeroen; Lordache, Marian-Daniel; Okujeni, Akpona; Hermy, Martin; van der Linden, Sebastian; Somers, Ben

2016-10-01

Remote sensing based land cover classification in urban areas generally requires the use of subpixel classification algorithms to take into account the high spatial heterogeneity. These spectral unmixing techniques often rely on spectral libraries, i.e. collections of pure material spectra (endmembers, EM), which ideally cover the large EM variability typically present in urban scenes. Despite the advent of several (semi-) automated EM detection algorithms, the collection of such image-specific libraries remains a tedious and time-consuming task. As an alternative, we suggest the use of a generic urban EM library, containing material spectra under varying conditions, acquired from different locations and sensors. This approach requires an efficient EM selection technique, capable of only selecting those spectra relevant for a specific image. In this paper, we evaluate and compare the potential of different existing library pruning algorithms (Iterative Endmember Selection and MUSIC) using simulated hyperspectral (APEX) data of the Brussels metropolitan area. In addition, we develop a new hybrid EM selection method which is shown to be highly efficient in dealing with both imagespecific and generic libraries, subsequently yielding more robust land cover classification results compared to existing methods. Future research will include further optimization of the proposed algorithm and additional tests on both simulated and real hyperspectral data.

SkICAT: A cataloging and analysis tool for wide field imaging surveys

NASA Technical Reports Server (NTRS)

Weir, N.; Fayyad, U. M.; Djorgovski, S. G.; Roden, J.

1992-01-01

We describe an integrated system, SkICAT (Sky Image Cataloging and Analysis Tool), for the automated reduction and analysis of the Palomar Observatory-ST ScI Digitized Sky Survey. The Survey will consist of the complete digitization of the photographic Second Palomar Observatory Sky Survey (POSS-II) in three bands, comprising nearly three Terabytes of pixel data. SkICAT applies a combination of existing packages, including FOCAS for basic image detection and measurement and SAS for database management, as well as custom software, to the task of managing this wealth of data. One of the most novel aspects of the system is its method of object classification. Using state-of-theart machine learning classification techniques (GID3* and O-BTree), we have developed a powerful method for automatically distinguishing point sources from non-point sources and artifacts, achieving comparably accurate discrimination a full magnitude fainter than in previous Schmidt plate surveys. The learning algorithms produce decision trees for classification by examining instances of objects classified by eye on both plate and higher quality CCD data. The same techniques will be applied to perform higher-level object classification (e.g., of galaxy morphology) in the near future. Another key feature of the system is the facility to integrate the catalogs from multiple plates (and portions thereof) to construct a single catalog of uniform calibration and quality down to the faintest limits of the survey. SkICAT also provides a variety of data analysis and exploration tools for the scientific utilization of the resulting catalogs. We include initial results of applying this system to measure the counts and distribution of galaxies in two bands down to Bj is approximately 21 mag over an approximate 70 square degree multi-plate field from POSS-II. SkICAT is constructed in a modular and general fashion and should be readily adaptable to other large-scale imaging surveys.

Pigeons (Columba livia) as Trainable Observers of Pathology and Radiology Breast Cancer Images.

PubMed

Levenson, Richard M; Krupinski, Elizabeth A; Navarro, Victor M; Wasserman, Edward A

2015-01-01

Pathologists and radiologists spend years acquiring and refining their medically essential visual skills, so it is of considerable interest to understand how this process actually unfolds and what image features and properties are critical for accurate diagnostic performance. Key insights into human behavioral tasks can often be obtained by using appropriate animal models. We report here that pigeons (Columba livia)-which share many visual system properties with humans-can serve as promising surrogate observers of medical images, a capability not previously documented. The birds proved to have a remarkable ability to distinguish benign from malignant human breast histopathology after training with differential food reinforcement; even more importantly, the pigeons were able to generalize what they had learned when confronted with novel image sets. The birds' histological accuracy, like that of humans, was modestly affected by the presence or absence of color as well as by degrees of image compression, but these impacts could be ameliorated with further training. Turning to radiology, the birds proved to be similarly capable of detecting cancer-relevant microcalcifications on mammogram images. However, when given a different (and for humans quite difficult) task-namely, classification of suspicious mammographic densities (masses)-the pigeons proved to be capable only of image memorization and were unable to successfully generalize when shown novel examples. The birds' successes and difficulties suggest that pigeons are well-suited to help us better understand human medical image perception, and may also prove useful in performance assessment and development of medical imaging hardware, image processing, and image analysis tools.

Categorization abilities for emotional and nonemotional stimuli in patients with alcohol-related Korsakoff syndrome.

PubMed

Labudda, Kirsten; von Rothkirch, Nadine; Pawlikowski, Mirko; Laier, Christian; Brand, Matthias

2010-06-01

To investigate whether patients with alcohol-related Korsakoff syndrome (KR) have emotion-specific or general deficits in multicategoric classification performance. Earlier studies have shown reduced performance in classifying stimuli according to their emotional valence in patients with KS. However, it is unclear whether such classification deficits are of emotion-specific nature or whether they can also occur when nonemotional classifications are demanded. In this study, we examined 35 patients with alcoholic KS and 35 healthy participants with the Emotional Picture Task (EPT) to assess valence classification performance, the Semantic Classification Task (SCT) to assess nonemotional categorizations, and an extensive neuropsychologic test battery. KS patients exhibited lower classification performance in both tasks compared with the healthy participants. EPT and SCT performance were related to each other. EPT and SCT performance correlated with general knowledge and EPT performance in addition with executive functions. Our results indicate a common underlying mechanism of the patients' reductions in emotional and nonemotional classification performance. These deficits are most probably based on problems in retrieving object and category knowledge and, partially, on executive functioning.

The EO-1 hyperion and advanced land imager sensors for use in tundra classification studies within the Upper Kuparuk River Basin, Alaska

NASA Astrophysics Data System (ADS)

Hall-Brown, Mary

The heterogeneity of Arctic vegetation can make land cover classification vey difficult when using medium to small resolution imagery (Schneider et al., 2009; Muller et al., 1999). Using high radiometric and spatial resolution imagery, such as the SPOT 5 and IKONOS satellites, have helped arctic land cover classification accuracies rise into the 80 and 90 percentiles (Allard, 2003; Stine et al., 2010; Muller et al., 1999). However, those increases usually come at a high price. High resolution imagery is very expensive and can often add tens of thousands of dollars onto the cost of the research. The EO-1 satellite launched in 2002 carries two sensors that have high specral and/or high spatial resolutions and can be an acceptable compromise between the resolution versus cost issues. The Hyperion is a hyperspectral sensor with the capability of collecting 242 spectral bands of information. The Advanced Land Imager (ALI) is an advanced multispectral sensor whose spatial resolution can be sharpened to 10 meters. This dissertation compares the accuracies of arctic land cover classifications produced by the Hyperion and ALI sensors to the classification accuracies produced by the Systeme Pour l' Observation de le Terre (SPOT), the Landsat Thematic Mapper (TM) and the Landsat Enhanced Thematic Mapper Plus (ETM+) sensors. Hyperion and ALI images from August 2004 were collected over the Upper Kuparuk River Basin, Alaska. Image processing included the stepwise discriminant analysis of pixels that were positively classified from coinciding ground control points, geometric and radiometric correction, and principle component analysis. Finally, stratified random sampling was used to perform accuracy assessments on satellite derived land cover classifications. Accuracy was estimated from an error matrix (confusion matrix) that provided the overall, producer's and user's accuracies. This research found that while the Hyperion sensor produced classfication accuracies that were equivalent to the TM and ETM+ sensor (approximately 78%), the Hyperion could not obtain the accuracy of the SPOT 5 HRV sensor. However, the land cover classifications derived from the ALI sensor exceeded most classification accuracies derived from the TM and ETM+ senors and were even comparable to most SPOT 5 HRV classifications (87%). With the deactivation of the Landsat series satellites, the monitoring of remote locations such as in the Arctic on an uninterupted basis thoughout the world is in jeopardy. The utilization of the Hyperion and ALI sensors are a way to keep that endeavor operational. By keeping the ALI sensor active at all times, uninterupted observation of the entire Earth can be accomplished. Keeping the Hyperion sensor as a "tasked" sensor can provide scientists with additional imagery and options for their studies without overburdening storage issues.

Land Cover Classification in a Complex Urban-Rural Landscape with Quickbird Imagery

PubMed Central

Moran, Emilio Federico.

2010-01-01

High spatial resolution images have been increasingly used for urban land use/cover classification, but the high spectral variation within the same land cover, the spectral confusion among different land covers, and the shadow problem often lead to poor classification performance based on the traditional per-pixel spectral-based classification methods. This paper explores approaches to improve urban land cover classification with Quickbird imagery. Traditional per-pixel spectral-based supervised classification, incorporation of textural images and multispectral images, spectral-spatial classifier, and segmentation-based classification are examined in a relatively new developing urban landscape, Lucas do Rio Verde in Mato Grosso State, Brazil. This research shows that use of spatial information during the image classification procedure, either through the integrated use of textural and spectral images or through the use of segmentation-based classification method, can significantly improve land cover classification performance. PMID:21643433

Integrating human and machine intelligence in galaxy morphology classification tasks

NASA Astrophysics Data System (ADS)

Beck, Melanie R.; Scarlata, Claudia; Fortson, Lucy F.; Lintott, Chris J.; Simmons, B. D.; Galloway, Melanie A.; Willett, Kyle W.; Dickinson, Hugh; Masters, Karen L.; Marshall, Philip J.; Wright, Darryl

2018-06-01

Quantifying galaxy morphology is a challenging yet scientifically rewarding task. As the scale of data continues to increase with upcoming surveys, traditional classification methods will struggle to handle the load. We present a solution through an integration of visual and automated classifications, preserving the best features of both human and machine. We demonstrate the effectiveness of such a system through a re-analysis of visual galaxy morphology classifications collected during the Galaxy Zoo 2 (GZ2) project. We reprocess the top-level question of the GZ2 decision tree with a Bayesian classification aggregation algorithm dubbed SWAP, originally developed for the Space Warps gravitational lens project. Through a simple binary classification scheme, we increase the classification rate nearly 5-fold classifying 226 124 galaxies in 92 d of GZ2 project time while reproducing labels derived from GZ2 classification data with 95.7 per cent accuracy. We next combine this with a Random Forest machine learning algorithm that learns on a suite of non-parametric morphology indicators widely used for automated morphologies. We develop a decision engine that delegates tasks between human and machine and demonstrate that the combined system provides at least a factor of 8 increase in the classification rate, classifying 210 803 galaxies in just 32 d of GZ2 project time with 93.1 per cent accuracy. As the Random Forest algorithm requires a minimal amount of computational cost, this result has important implications for galaxy morphology identification tasks in the era of Euclid and other large-scale surveys.

Iris Image Classification Based on Hierarchical Visual Codebook.

PubMed

Zhenan Sun; Hui Zhang; Tieniu Tan; Jianyu Wang

2014-06-01

Iris recognition as a reliable method for personal identification has been well-studied with the objective to assign the class label of each iris image to a unique subject. In contrast, iris image classification aims to classify an iris image to an application specific category, e.g., iris liveness detection (classification of genuine and fake iris images), race classification (e.g., classification of iris images of Asian and non-Asian subjects), coarse-to-fine iris identification (classification of all iris images in the central database into multiple categories). This paper proposes a general framework for iris image classification based on texture analysis. A novel texture pattern representation method called Hierarchical Visual Codebook (HVC) is proposed to encode the texture primitives of iris images. The proposed HVC method is an integration of two existing Bag-of-Words models, namely Vocabulary Tree (VT), and Locality-constrained Linear Coding (LLC). The HVC adopts a coarse-to-fine visual coding strategy and takes advantages of both VT and LLC for accurate and sparse representation of iris texture. Extensive experimental results demonstrate that the proposed iris image classification method achieves state-of-the-art performance for iris liveness detection, race classification, and coarse-to-fine iris identification. A comprehensive fake iris image database simulating four types of iris spoof attacks is developed as the benchmark for research of iris liveness detection.

Object Classification in Semi Structured Enviroment Using Forward-Looking Sonar

PubMed Central

dos Santos, Matheus; Ribeiro, Pedro Otávio; Núñez, Pedro; Botelho, Silvia

2017-01-01

The submarine exploration using robots has been increasing in recent years. The automation of tasks such as monitoring, inspection, and underwater maintenance requires the understanding of the robot’s environment. The object recognition in the scene is becoming a critical issue for these systems. On this work, an underwater object classification pipeline applied in acoustic images acquired by Forward-Looking Sonar (FLS) are studied. The object segmentation combines thresholding, connected pixels searching and peak of intensity analyzing techniques. The object descriptor extract intensity and geometric features of the detected objects. A comparison between the Support Vector Machine, K-Nearest Neighbors, and Random Trees classifiers are presented. An open-source tool was developed to annotate and classify the objects and evaluate their classification performance. The proposed method efficiently segments and classifies the structures in the scene using a real dataset acquired by an underwater vehicle in a harbor area. Experimental results demonstrate the robustness and accuracy of the method described in this paper. PMID:28961163

Accuracy of automated classification of major depressive disorder as a function of symptom severity.

PubMed

Ramasubbu, Rajamannar; Brown, Matthew R G; Cortese, Filmeno; Gaxiola, Ismael; Goodyear, Bradley; Greenshaw, Andrew J; Dursun, Serdar M; Greiner, Russell

2016-01-01

Growing evidence documents the potential of machine learning for developing brain based diagnostic methods for major depressive disorder (MDD). As symptom severity may influence brain activity, we investigated whether the severity of MDD affected the accuracies of machine learned MDD-vs-Control diagnostic classifiers. Forty-five medication-free patients with DSM-IV defined MDD and 19 healthy controls participated in the study. Based on depression severity as determined by the Hamilton Rating Scale for Depression (HRSD), MDD patients were sorted into three groups: mild to moderate depression (HRSD 14-19), severe depression (HRSD 20-23), and very severe depression (HRSD ≥ 24). We collected functional magnetic resonance imaging (fMRI) data during both resting-state and an emotional-face matching task. Patients in each of the three severity groups were compared against controls in separate analyses, using either the resting-state or task-based fMRI data. We use each of these six datasets with linear support vector machine (SVM) binary classifiers for identifying individuals as patients or controls. The resting-state fMRI data showed statistically significant classification accuracy only for the very severe depression group (accuracy 66%, p = 0.012 corrected), while mild to moderate (accuracy 58%, p = 1.0 corrected) and severe depression (accuracy 52%, p = 1.0 corrected) were only at chance. With task-based fMRI data, the automated classifier performed at chance in all three severity groups. Binary linear SVM classifiers achieved significant classification of very severe depression with resting-state fMRI, but the contribution of brain measurements may have limited potential in differentiating patients with less severe depression from healthy controls.

Lossless Compression of Classification-Map Data

NASA Technical Reports Server (NTRS)

Hua, Xie; Klimesh, Matthew

2009-01-01

A lossless image-data-compression algorithm intended specifically for application to classification-map data is based on prediction, context modeling, and entropy coding. The algorithm was formulated, in consideration of the differences between classification maps and ordinary images of natural scenes, so as to be capable of compressing classification- map data more effectively than do general-purpose image-data-compression algorithms. Classification maps are typically generated from remote-sensing images acquired by instruments aboard aircraft (see figure) and spacecraft. A classification map is a synthetic image that summarizes information derived from one or more original remote-sensing image(s) of a scene. The value assigned to each pixel in such a map is the index of a class that represents some type of content deduced from the original image data for example, a type of vegetation, a mineral, or a body of water at the corresponding location in the scene. When classification maps are generated onboard the aircraft or spacecraft, it is desirable to compress the classification-map data in order to reduce the volume of data that must be transmitted to a ground station.

Medical image classification based on multi-scale non-negative sparse coding.

PubMed

Zhang, Ruijie; Shen, Jian; Wei, Fushan; Li, Xiong; Sangaiah, Arun Kumar

2017-11-01

With the rapid development of modern medical imaging technology, medical image classification has become more and more important in medical diagnosis and clinical practice. Conventional medical image classification algorithms usually neglect the semantic gap problem between low-level features and high-level image semantic, which will largely degrade the classification performance. To solve this problem, we propose a multi-scale non-negative sparse coding based medical image classification algorithm. Firstly, Medical images are decomposed into multiple scale layers, thus diverse visual details can be extracted from different scale layers. Secondly, for each scale layer, the non-negative sparse coding model with fisher discriminative analysis is constructed to obtain the discriminative sparse representation of medical images. Then, the obtained multi-scale non-negative sparse coding features are combined to form a multi-scale feature histogram as the final representation for a medical image. Finally, SVM classifier is combined to conduct medical image classification. The experimental results demonstrate that our proposed algorithm can effectively utilize multi-scale and contextual spatial information of medical images, reduce the semantic gap in a large degree and improve medical image classification performance. Copyright © 2017 Elsevier B.V. All rights reserved.

Retinex Preprocessing for Improved Multi-Spectral Image Classification

NASA Technical Reports Server (NTRS)

Thompson, B.; Rahman, Z.; Park, S.

2000-01-01

The goal of multi-image classification is to identify and label "similar regions" within a scene. The ability to correctly classify a remotely sensed multi-image of a scene is affected by the ability of the classification process to adequately compensate for the effects of atmospheric variations and sensor anomalies. Better classification may be obtained if the multi-image is preprocessed before classification, so as to reduce the adverse effects of image formation. In this paper, we discuss the overall impact on multi-spectral image classification when the retinex image enhancement algorithm is used to preprocess multi-spectral images. The retinex is a multi-purpose image enhancement algorithm that performs dynamic range compression, reduces the dependence on lighting conditions, and generally enhances apparent spatial resolution. The retinex has been successfully applied to the enhancement of many different types of grayscale and color images. We show in this paper that retinex preprocessing improves the spatial structure of multi-spectral images and thus provides better within-class variations than would otherwise be obtained without the preprocessing. For a series of multi-spectral images obtained with diffuse and direct lighting, we show that without retinex preprocessing the class spectral signatures vary substantially with the lighting conditions. Whereas multi-dimensional clustering without preprocessing produced one-class homogeneous regions, the classification on the preprocessed images produced multi-class non-homogeneous regions. This lack of homogeneity is explained by the interaction between different agronomic treatments applied to the regions: the preprocessed images are closer to ground truth. The principle advantage that the retinex offers is that for different lighting conditions classifications derived from the retinex preprocessed images look remarkably "similar", and thus more consistent, whereas classifications derived from the original images, without preprocessing, are much less similar.

Working memory supports inference learning just like classification learning.

PubMed

Craig, Stewart; Lewandowsky, Stephan

2013-08-01

Recent research has found a positive relationship between people's working memory capacity (WMC) and their speed of category learning. To date, only classification-learning tasks have been considered, in which people learn to assign category labels to objects. It is unknown whether learning to make inferences about category features might also be related to WMC. We report data from a study in which 119 participants undertook classification learning and inference learning, and completed a series of WMC tasks. Working memory capacity was positively related to people's classification and inference learning performance.

Typicality effects in artificial categories: is there a hemisphere difference?

PubMed

Richards, L G; Chiarello, C

1990-07-01

In category classification tasks, typicality effects are usually found: accuracy and reaction time depend upon distance from a prototype. In this study, subjects learned either verbal or nonverbal dot pattern categories, followed by a lateralized classification task. Comparable typicality effects were found in both reaction time and accuracy across visual fields for both verbal and nonverbal categories. Both hemispheres appeared to use a similarity-to-prototype matching strategy in classification. This indicates that merely having a verbal label does not differentiate classification in the two hemispheres.

Morphological classification of odontogenic keratocysts using Bouligand-Minkowski fractal descriptors.

PubMed

Florindo, Joao B; Bruno, Odemir M; Landini, Gabriel

2017-02-01

The Odontogenic keratocyst (OKC) is a cystic lesion of the jaws, which has high growth and recurrence rates compared to other cysts of the jaws (for instance, radicular cyst, which is the most common jaw cyst type). For this reason OKCs are considered by some to be benign neoplasms. There exist two sub-types of OKCs (sporadic and syndromic) and the ability to discriminate between these sub-types, as well as other jaw cysts, is an important task in terms of disease diagnosis and prognosis. With the development of digital pathology, computational algorithms have become central to addressing this type of problem. Considering that only basic feature-based methods have been investigated in this problem before, we propose to use a different approach (the Bouligand-Minkowski descriptors) to assess the success rates achieved on the classification of a database of histological images of the epithelial lining of these cysts. This does not require the level of abstraction necessary to extract histologically-relevant features and therefore has the potential of being more robust than previous approaches. The descriptors were obtained by mapping pixel intensities into a three dimensional cloud of points in discrete space and applying morphological dilations with spheres of increasing radii. The descriptors were computed from the volume of the dilated set and submitted to a machine learning algorithm to classify the samples into diagnostic groups. This approach was capable of discriminating between OKCs and radicular cysts in 98% of images (100% of cases) and between the two sub-types of OKCs in 68% of images (71% of cases). These results improve over previously reported classification rates reported elsewhere and suggest that Bouligand-Minkowski descriptors are useful features to be used in histopathological images of these cysts. Copyright © 2016 The Authors. Published by Elsevier Ltd.. All rights reserved.

Volumetric quantitative characterization of human patellar cartilage with topological and geometrical features on phase-contrast X-ray computed tomography.

PubMed

Nagarajan, Mahesh B; Coan, Paola; Huber, Markus B; Diemoz, Paul C; Wismüller, Axel

2015-11-01

Phase-contrast X-ray computed tomography (PCI-CT) has attracted significant interest in recent years for its ability to provide significantly improved image contrast in low absorbing materials such as soft biological tissue. In the research context of cartilage imaging, previous studies have demonstrated the ability of PCI-CT to visualize structural details of human patellar cartilage matrix and capture changes to chondrocyte organization induced by osteoarthritis. This study evaluates the use of geometrical and topological features for volumetric characterization of such chondrocyte patterns in the presence (or absence) of osteoarthritic damage. Geometrical features derived from the scaling index method (SIM) and topological features derived from Minkowski Functionals were extracted from 1392 volumes of interest (VOI) annotated on PCI-CT images of ex vivo human patellar cartilage specimens. These features were subsequently used in a machine learning task with support vector regression to classify VOIs as healthy or osteoarthritic; classification performance was evaluated using the area under the receiver operating characteristic curve (AUC). Our results show that the classification performance of SIM-derived geometrical features (AUC: 0.90 ± 0.09) is significantly better than Minkowski Functionals volume (AUC: 0.54 ± 0.02), surface (AUC: 0.72 ± 0.06), mean breadth (AUC: 0.74 ± 0.06) and Euler characteristic (AUC: 0.78 ± 0.04) (p < 10(-4)). These results suggest that such geometrical features can provide a detailed characterization of the chondrocyte organization in the cartilage matrix in an automated manner, while also enabling classification of cartilage as healthy or osteoarthritic with high accuracy. Such features could potentially serve as diagnostic imaging markers for evaluating osteoarthritis progression and its response to different therapeutic intervention strategies.

Performance Evaluation of Frequency Transform Based Block Classification of Compound Image Segmentation Techniques

NASA Astrophysics Data System (ADS)

Selwyn, Ebenezer Juliet; Florinabel, D. Jemi

2018-04-01

Compound image segmentation plays a vital role in the compression of computer screen images. Computer screen images are images which are mixed with textual, graphical, or pictorial contents. In this paper, we present a comparison of two transform based block classification of compound images based on metrics like speed of classification, precision and recall rate. Block based classification approaches normally divide the compound images into fixed size blocks of non-overlapping in nature. Then frequency transform like Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) are applied over each block. Mean and standard deviation are computed for each 8 × 8 block and are used as features set to classify the compound images into text/graphics and picture/background block. The classification accuracy of block classification based segmentation techniques are measured by evaluation metrics like precision and recall rate. Compound images of smooth background and complex background images containing text of varying size, colour and orientation are considered for testing. Experimental evidence shows that the DWT based segmentation provides significant improvement in recall rate and precision rate approximately 2.3% than DCT based segmentation with an increase in block classification time for both smooth and complex background images.

Crowdsourcing as a screening tool to detect clinical features of glaucomatous optic neuropathy from digital photography.

PubMed

Mitry, Danny; Peto, Tunde; Hayat, Shabina; Blows, Peter; Morgan, James; Khaw, Kay-Tee; Foster, Paul J

2015-01-01

Crowdsourcing is the process of simplifying and outsourcing numerous tasks to many untrained individuals. Our aim was to assess the performance and repeatability of crowdsourcing in the classification of normal and glaucomatous discs from optic disc images. Optic disc images (N = 127) with pre-determined disease status were selected by consensus agreement from grading experts from a large cohort study. After reading brief illustrative instructions, we requested that knowledge workers (KWs) from a crowdsourcing platform (Amazon MTurk) classified each image as normal or abnormal. Each image was classified 20 times by different KWs. Two study designs were examined to assess the effect of varying KW experience and both study designs were conducted twice for consistency. Performance was assessed by comparing the sensitivity, specificity and area under the receiver operating characteristic curve (AUC). Overall, 2,540 classifications were received in under 24 hours at minimal cost. The sensitivity ranged between 83-88% across both trials and study designs, however the specificity was poor, ranging between 35-43%. In trial 1, the highest AUC (95%CI) was 0.64(0.62-0.66) and in trial 2 it was 0.63(0.61-0.65). There were no significant differences between study design or trials conducted. Crowdsourcing represents a cost-effective method of image analysis which demonstrates good repeatability and a high sensitivity. Optimisation of variables such as reward schemes, mode of image presentation, expanded response options and incorporation of training modules should be examined to determine their effect on the accuracy and reliability of this technique in retinal image analysis.

Developmental Inter-Relationships Among Concrete Operational Tasks: An Investigation of Piaget's Stage Concept

ERIC Educational Resources Information Center

Jamison, Wesley

1977-01-01

Two models of intertask relations, Wohlwill's divergent-decalage and reciprocal-interaction patterns, were evaluated for their fit to cross-classification tables which showed the joint classification of 101 children's performance on all possible pairs of eight concrete operational tasks. (SB)

Compression of deep convolutional neural network for computer-aided diagnosis of masses in digital breast tomosynthesis

NASA Astrophysics Data System (ADS)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir; Helvie, Mark A.; Richter, Caleb; Cha, Kenny

2018-02-01

Deep-learning models are highly parameterized, causing difficulty in inference and transfer learning. We propose a layered pathway evolution method to compress a deep convolutional neural network (DCNN) for classification of masses in DBT while maintaining the classification accuracy. Two-stage transfer learning was used to adapt the ImageNet-trained DCNN to mammography and then to DBT. In the first-stage transfer learning, transfer learning from ImageNet trained DCNN was performed using mammography data. In the second-stage transfer learning, the mammography-trained DCNN was trained on the DBT data using feature extraction from fully connected layer, recursive feature elimination and random forest classification. The layered pathway evolution encapsulates the feature extraction to the classification stages to compress the DCNN. Genetic algorithm was used in an iterative approach with tournament selection driven by count-preserving crossover and mutation to identify the necessary nodes in each convolution layer while eliminating the redundant nodes. The DCNN was reduced by 99% in the number of parameters and 95% in mathematical operations in the convolutional layers. The lesion-based area under the receiver operating characteristic curve on an independent DBT test set from the original and the compressed network resulted in 0.88+/-0.05 and 0.90+/-0.04, respectively. The difference did not reach statistical significance. We demonstrated a DCNN compression approach without additional fine-tuning or loss of performance for classification of masses in DBT. The approach can be extended to other DCNNs and transfer learning tasks. An ensemble of these smaller and focused DCNNs has the potential to be used in multi-target transfer learning.

Resolving task rule incongruence during task switching by competitor rule suppression.

PubMed

Meiran, Nachshon; Hsieh, Shulan; Dimov, Eduard

2010-07-01

Task switching requires maintaining readiness to execute any task of a given set of tasks. However, when tasks switch, the readiness to execute the now-irrelevant task generates interference, as seen in the task rule incongruence effect. Overcoming such interference requires fine-tuned inhibition that impairs task readiness only minimally. In an experiment involving 2 object classification tasks and 2 location classification tasks, the authors show that irrelevant task rules that generate response conflicts are inhibited. This competitor rule suppression (CRS) is seen in response slowing in subsequent trials, when the competing rules become relevant. CRS is shown to operate on specific rules without affecting similar rules. CRS and backward inhibition, which is another inhibitory phenomenon, produced additive effects on reaction time, suggesting their mutual independence. Implications for current formal theories of task switching as well as for conflict monitoring theories are discussed. (c) 2010 APA, all rights reserved

Agricultural Land Use mapping by multi-sensor approach for hydrological water quality monitoring

NASA Astrophysics Data System (ADS)

Brodsky, Lukas; Kodesova, Radka; Kodes, Vit

2010-05-01

The main objective of this study is to demonstrate potential of operational use of the high and medium resolution remote sensing data for hydrological water quality monitoring by mapping agriculture intensity and crop structures. In particular use of remote sensing mapping for optimization of pesticide monitoring. The agricultural mapping task is tackled by means of medium spatial and high temporal resolution ESA Envisat MERIS FR images together with single high spatial resolution IRS AWiFS image covering the whole area of interest (the Czech Republic). High resolution data (e.g. SPOT, ALOS, Landsat) are often used for agricultural land use classification, but usually only at regional or local level due to data availability and financial constraints. AWiFS data (nominal spatial resolution 56 m) due to the wide satellite swath seems to be more suitable for use at national level. Nevertheless, one of the critical issues for such a classification is to have sufficient image acquisitions over the whole vegetation period to describe crop development in appropriate way. ESA MERIS middle-resolution data were used in several studies for crop classification. The high temporal and also spectral resolution of MERIS data has indisputable advantage for crop classification. However, spatial resolution of 300 m results in mixture signal in a single pixel. AWiFS-MERIS data synergy brings new perspectives in agricultural Land Use mapping. Also, the developed methodology procedure is fully compatible with future use of ESA (GMES) Sentinel satellite images. The applied methodology of hybrid multi-sensor approach consists of these main stages: a/ parcel segmentation and spectral pre-classification of high resolution image (AWiFS); b/ ingestion of middle resolution (MERIS) vegetation spectro-temporal features; c/ vegetation signatures unmixing; and d/ semantic object-oriented classification of vegetation classes into final classification scheme. These crop groups were selected to be classified: winter crops, spring crops, oilseed rape, legumes, summer and other crops. This study highlights operational potentials of high temporal full resolution MERIS images in agricultural land use monitoring. Practical application of this methodology is foreseen, among others, in the water quality monitoring. Effective pesticide monitoring relies also on spatial distribution of applied pesticides, which can be derived from crop - plant protection product relationship. Knowledge of areas with predominant occurrence of specific crop based on remote sensing data described above can be used for a forecast of probable plant protection product application, thus cost-effective pesticide monitoring. The remote sensing data used on a continuous basis can be used in other long-term water management issues and provide valuable data for decision makers. Acknowledgement: Authors acknowledge the financial support of the Ministry of Education, Youth and Sports of the Czech Republic (grants No. 2B06095 and No. MSM 6046070901). The study was also supported by ESA CAT-1 (ref. 4358) and SOSI projects (Spatial Observation Services and Infrastructure; ref. GSTP-RTDA-EOPG-SW-08-0004).

Segmentation of white blood cells and comparison of cell morphology by linear and naïve Bayes classifiers.

PubMed

Prinyakupt, Jaroonrut; Pluempitiwiriyawej, Charnchai

2015-06-30

Blood smear microscopic images are routinely investigated by haematologists to diagnose most blood diseases. However, the task is quite tedious and time consuming. An automatic detection and classification of white blood cells within such images can accelerate the process tremendously. In this paper we propose a system to locate white blood cells within microscopic blood smear images, segment them into nucleus and cytoplasm regions, extract suitable features and finally, classify them into five types: basophil, eosinophil, neutrophil, lymphocyte and monocyte. Two sets of blood smear images were used in this study's experiments. Dataset 1, collected from Rangsit University, were normal peripheral blood slides under light microscope with 100× magnification; 555 images with 601 white blood cells were captured by a Nikon DS-Fi2 high-definition color camera and saved in JPG format of size 960 × 1,280 pixels at 15 pixels per 1 μm resolution. In dataset 2, 477 cropped white blood cell images were downloaded from CellaVision.com. They are in JPG format of size 360 × 363 pixels. The resolution is estimated to be 10 pixels per 1 μm. The proposed system comprises a pre-processing step, nucleus segmentation, cell segmentation, feature extraction, feature selection and classification. The main concept of the segmentation algorithm employed uses white blood cell's morphological properties and the calibrated size of a real cell relative to image resolution. The segmentation process combined thresholding, morphological operation and ellipse curve fitting. Consequently, several features were extracted from the segmented nucleus and cytoplasm regions. Prominent features were then chosen by a greedy search algorithm called sequential forward selection. Finally, with a set of selected prominent features, both linear and naïve Bayes classifiers were applied for performance comparison. This system was tested on normal peripheral blood smear slide images from two datasets. Two sets of comparison were performed: segmentation and classification. The automatically segmented results were compared to the ones obtained manually by a haematologist. It was found that the proposed method is consistent and coherent in both datasets, with dice similarity of 98.9 and 91.6% for average segmented nucleus and cell regions, respectively. Furthermore, the overall correction rate in the classification phase is about 98 and 94% for linear and naïve Bayes models, respectively. The proposed system, based on normal white blood cell morphology and its characteristics, was applied to two different datasets. The results of the calibrated segmentation process on both datasets are fast, robust, efficient and coherent. Meanwhile, the classification of normal white blood cells into five types shows high sensitivity in both linear and naïve Bayes models, with slightly better results in the linear classifier.

Unmanned Aerial Vehicles (UAVs) and Artificial Intelligence Revolutionizing Wildlife Monitoring and Conservation

PubMed Central

Gonzalez, Luis F.; Montes, Glen A.; Puig, Eduard; Johnson, Sandra; Mengersen, Kerrie; Gaston, Kevin J.

2016-01-01

Surveying threatened and invasive species to obtain accurate population estimates is an important but challenging task that requires a considerable investment in time and resources. Estimates using existing ground-based monitoring techniques, such as camera traps and surveys performed on foot, are known to be resource intensive, potentially inaccurate and imprecise, and difficult to validate. Recent developments in unmanned aerial vehicles (UAV), artificial intelligence and miniaturized thermal imaging systems represent a new opportunity for wildlife experts to inexpensively survey relatively large areas. The system presented in this paper includes thermal image acquisition as well as a video processing pipeline to perform object detection, classification and tracking of wildlife in forest or open areas. The system is tested on thermal video data from ground based and test flight footage, and is found to be able to detect all the target wildlife located in the surveyed area. The system is flexible in that the user can readily define the types of objects to classify and the object characteristics that should be considered during classification. PMID:26784196

Unmanned Aerial Vehicles (UAVs) and Artificial Intelligence Revolutionizing Wildlife Monitoring and Conservation.

PubMed

Gonzalez, Luis F; Montes, Glen A; Puig, Eduard; Johnson, Sandra; Mengersen, Kerrie; Gaston, Kevin J

2016-01-14

Surveying threatened and invasive species to obtain accurate population estimates is an important but challenging task that requires a considerable investment in time and resources. Estimates using existing ground-based monitoring techniques, such as camera traps and surveys performed on foot, are known to be resource intensive, potentially inaccurate and imprecise, and difficult to validate. Recent developments in unmanned aerial vehicles (UAV), artificial intelligence and miniaturized thermal imaging systems represent a new opportunity for wildlife experts to inexpensively survey relatively large areas. The system presented in this paper includes thermal image acquisition as well as a video processing pipeline to perform object detection, classification and tracking of wildlife in forest or open areas. The system is tested on thermal video data from ground based and test flight footage, and is found to be able to detect all the target wildlife located in the surveyed area. The system is flexible in that the user can readily define the types of objects to classify and the object characteristics that should be considered during classification.

Using spectral imaging for the analysis of abnormalities for colorectal cancer: When is it helpful?

PubMed

Awan, Ruqayya; Al-Maadeed, Somaya; Al-Saady, Rafif

2018-01-01

The spectral imaging technique has been shown to provide more discriminative information than the RGB images and has been proposed for a range of problems. There are many studies demonstrating its potential for the analysis of histopathology images for abnormality detection but there have been discrepancies among previous studies as well. Many multispectral based methods have been proposed for histopathology images but the significance of the use of whole multispectral cube versus a subset of bands or a single band is still arguable. We performed comprehensive analysis using individual bands and different subsets of bands to determine the effectiveness of spectral information for determining the anomaly in colorectal images. Our multispectral colorectal dataset consists of four classes, each represented by infra-red spectrum bands in addition to the visual spectrum bands. We performed our analysis of spectral imaging by stratifying the abnormalities using both spatial and spectral information. For our experiments, we used a combination of texture descriptors with an ensemble classification approach that performed best on our dataset. We applied our method to another dataset and got comparable results with those obtained using the state-of-the-art method and convolutional neural network based method. Moreover, we explored the relationship of the number of bands with the problem complexity and found that higher number of bands is required for a complex task to achieve improved performance. Our results demonstrate a synergy between infra-red and visual spectrum by improving the classification accuracy (by 6%) on incorporating the infra-red representation. We also highlight the importance of how the dataset should be divided into training and testing set for evaluating the histopathology image-based approaches, which has not been considered in previous studies on multispectral histopathology images.

Using spectral imaging for the analysis of abnormalities for colorectal cancer: When is it helpful?

PubMed Central

Al-Maadeed, Somaya; Al-Saady, Rafif

2018-01-01

The spectral imaging technique has been shown to provide more discriminative information than the RGB images and has been proposed for a range of problems. There are many studies demonstrating its potential for the analysis of histopathology images for abnormality detection but there have been discrepancies among previous studies as well. Many multispectral based methods have been proposed for histopathology images but the significance of the use of whole multispectral cube versus a subset of bands or a single band is still arguable. We performed comprehensive analysis using individual bands and different subsets of bands to determine the effectiveness of spectral information for determining the anomaly in colorectal images. Our multispectral colorectal dataset consists of four classes, each represented by infra-red spectrum bands in addition to the visual spectrum bands. We performed our analysis of spectral imaging by stratifying the abnormalities using both spatial and spectral information. For our experiments, we used a combination of texture descriptors with an ensemble classification approach that performed best on our dataset. We applied our method to another dataset and got comparable results with those obtained using the state-of-the-art method and convolutional neural network based method. Moreover, we explored the relationship of the number of bands with the problem complexity and found that higher number of bands is required for a complex task to achieve improved performance. Our results demonstrate a synergy between infra-red and visual spectrum by improving the classification accuracy (by 6%) on incorporating the infra-red representation. We also highlight the importance of how the dataset should be divided into training and testing set for evaluating the histopathology image-based approaches, which has not been considered in previous studies on multispectral histopathology images. PMID:29874262

ICA-Based Imagined Conceptual Words Classification on EEG Signals.

PubMed

Imani, Ehsan; Pourmohammad, Ali; Bagheri, Mahsa; Mobasheri, Vida

2017-01-01

Independent component analysis (ICA) has been used for detecting and removing the eye artifacts conventionally. However, in this research, it was used not only for detecting the eye artifacts, but also for detecting the brain-produced signals of two conceptual danger and information category words. In this cross-sectional research, electroencephalography (EEG) signals were recorded using Micromed and 19-channel helmet devices in unipolar mode, wherein Cz electrode was selected as the reference electrode. In the first part of this research, the statistical community test case included four men and four women, who were 25-30 years old. In the designed task, three groups of traffic signs were considered, in which two groups referred to the concept of danger, and the third one referred to the concept of information. In the second part, the three volunteers, two men and one woman, who had the best results, were chosen from among eight participants. In the second designed task, direction arrows (up, down, left, and right) were used. For the 2/8 volunteers in the rest times, very high-power alpha waves were observed from the back of the head; however, in the thinking times, they were different. According to this result, alpha waves for changing the task from thinking to rest condition took at least 3 s for the two volunteers, and it was at most 5 s until they went to the absolute rest condition. For the 7/8 volunteers, the danger and information signals were well classified; these differences for the 5/8 volunteers were observed in the right hemisphere, and, for the other three volunteers, the differences were observed in the left hemisphere. For the second task, simulations showed that the best classification accuracies resulted when the time window was 2.5 s. In addition, it also showed that the features of the autoregressive (AR)-15 model coefficients were the best choices for extracting the features. For all the states of neural network except hardlim discriminator function, the classification accuracies were almost the same and not very different. Linear discriminant analysis (LDA) in comparison with the neural network yielded higher classification accuracies. ICA is a suitable algorithm for recognizing of the word's concept and its place in the brain. Achieved results from this experiment were the same compared with the results from other methods such as functional magnetic resonance imaging and methods based on the brain signals (EEG) in the vowel imagination and covert speech. Herein, the highest classification accuracy was obtained by extracting the target signal from the output of the ICA and extracting the features of coefficients AR model with time interval of 2.5 s. Finally, LDA resulted in the highest classification accuracy more than 60%.

Adaptive illumination source for multispectral vision system applied to material discrimination

NASA Astrophysics Data System (ADS)

Conde, Olga M.; Cobo, Adolfo; Cantero, Paulino; Conde, David; Mirapeix, Jesús; Cubillas, Ana M.; López-Higuera, José M.

2008-04-01

A multispectral system based on a monochrome camera and an adaptive illumination source is presented in this paper. Its preliminary application is focused on material discrimination for food and beverage industries, where monochrome, color and infrared imaging have been successfully applied for this task. This work proposes a different approach, in which the relevant wavelengths for the required discrimination task are selected in advance using a Sequential Forward Floating Selection (SFFS) Algorithm. A light source, based on Light Emitting Diodes (LEDs) at these wavelengths is then used to sequentially illuminate the material under analysis, and the resulting images are captured by a CCD camera with spectral response in the entire range of the selected wavelengths. Finally, the several multispectral planes obtained are processed using a Spectral Angle Mapping (SAM) algorithm, whose output is the desired material classification. Among other advantages, this approach of controlled and specific illumination produces multispectral imaging with a simple monochrome camera, and cold illumination restricted to specific relevant wavelengths, which is desirable for the food and beverage industry. The proposed system has been tested with success for the automatic detection of foreign object in the tobacco processing industry.

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.

Sentiment classification technology based on Markov logic networks

NASA Astrophysics Data System (ADS)

He, Hui; Li, Zhigang; Yao, Chongchong; Zhang, Weizhe

2016-07-01

With diverse online media emerging, there is a growing concern of sentiment classification problem. At present, text sentiment classification mainly utilizes supervised machine learning methods, which feature certain domain dependency. On the basis of Markov logic networks (MLNs), this study proposed a cross-domain multi-task text sentiment classification method rooted in transfer learning. Through many-to-one knowledge transfer, labeled text sentiment classification, knowledge was successfully transferred into other domains, and the precision of the sentiment classification analysis in the text tendency domain was improved. The experimental results revealed the following: (1) the model based on a MLN demonstrated higher precision than the single individual learning plan model. (2) Multi-task transfer learning based on Markov logical networks could acquire more knowledge than self-domain learning. The cross-domain text sentiment classification model could significantly improve the precision and efficiency of text sentiment classification.

Crown-level tree species classification from AISA hyperspectral imagery using an innovative pixel-weighting approach

NASA Astrophysics Data System (ADS)

Liu, Haijian; Wu, Changshan

2018-06-01

Crown-level tree species classification is a challenging task due to the spectral similarity among different tree species. Shadow, underlying objects, and other materials within a crown may decrease the purity of extracted crown spectra and further reduce classification accuracy. To address this problem, an innovative pixel-weighting approach was developed for tree species classification at the crown level. The method utilized high density discrete LiDAR data for individual tree delineation and Airborne Imaging Spectrometer for Applications (AISA) hyperspectral imagery for pure crown-scale spectra extraction. Specifically, three steps were included: 1) individual tree identification using LiDAR data, 2) pixel-weighted representative crown spectra calculation using hyperspectral imagery, with which pixel-based illuminated-leaf fractions estimated using a linear spectral mixture analysis (LSMA) were employed as weighted factors, and 3) representative spectra based tree species classification was performed through applying a support vector machine (SVM) approach. Analysis of results suggests that the developed pixel-weighting approach (OA = 82.12%, Kc = 0.74) performed better than treetop-based (OA = 70.86%, Kc = 0.58) and pixel-majority methods (OA = 72.26, Kc = 0.62) in terms of classification accuracy. McNemar tests indicated the differences in accuracy between pixel-weighting and treetop-based approaches as well as that between pixel-weighting and pixel-majority approaches were statistically significant.

Semantic and topological classification of images in magnetically guided capsule endoscopy

NASA Astrophysics Data System (ADS)

Mewes, P. W.; Rennert, P.; Juloski, A. L.; Lalande, A.; Angelopoulou, E.; Kuth, R.; Hornegger, J.

2012-03-01

Magnetically-guided capsule endoscopy (MGCE) is a nascent technology with the goal to allow the steering of a capsule endoscope inside a water filled stomach through an external magnetic field. We developed a classification cascade for MGCE images with groups images in semantic and topological categories. Results can be used in a post-procedure review or as a starting point for algorithms classifying pathologies. The first semantic classification step discards over-/under-exposed images as well as images with a large amount of debris. The second topological classification step groups images with respect to their position in the upper gastrointestinal tract (mouth, esophagus, stomach, duodenum). In the third stage two parallel classifications steps distinguish topologically different regions inside the stomach (cardia, fundus, pylorus, antrum, peristaltic view). For image classification, global image features and local texture features were applied and their performance was evaluated. We show that the third classification step can be improved by a bubble and debris segmentation because it limits feature extraction to discriminative areas only. We also investigated the impact of segmenting intestinal folds on the identification of different semantic camera positions. The results of classifications with a support-vector-machine show the significance of color histogram features for the classification of corrupted images (97%). Features extracted from intestinal fold segmentation lead only to a minor improvement (3%) in discriminating different camera positions.

Continuous robust sound event classification using time-frequency features and deep learning

PubMed Central

Song, Yan; Xiao, Wei; Phan, Huy

2017-01-01

The automatic detection and recognition of sound events by computers is a requirement for a number of emerging sensing and human computer interaction technologies. Recent advances in this field have been achieved by machine learning classifiers working in conjunction with time-frequency feature representations. This combination has achieved excellent accuracy for classification of discrete sounds. The ability to recognise sounds under real-world noisy conditions, called robust sound event classification, is an especially challenging task that has attracted recent research attention. Another aspect of real-word conditions is the classification of continuous, occluded or overlapping sounds, rather than classification of short isolated sound recordings. This paper addresses the classification of noise-corrupted, occluded, overlapped, continuous sound recordings. It first proposes a standard evaluation task for such sounds based upon a common existing method for evaluating isolated sound classification. It then benchmarks several high performing isolated sound classifiers to operate with continuous sound data by incorporating an energy-based event detection front end. Results are reported for each tested system using the new task, to provide the first analysis of their performance for continuous sound event detection. In addition it proposes and evaluates a novel Bayesian-inspired front end for the segmentation and detection of continuous sound recordings prior to classification. PMID:28892478

Continuous robust sound event classification using time-frequency features and deep learning.

PubMed

McLoughlin, Ian; Zhang, Haomin; Xie, Zhipeng; Song, Yan; Xiao, Wei; Phan, Huy

2017-01-01

The automatic detection and recognition of sound events by computers is a requirement for a number of emerging sensing and human computer interaction technologies. Recent advances in this field have been achieved by machine learning classifiers working in conjunction with time-frequency feature representations. This combination has achieved excellent accuracy for classification of discrete sounds. The ability to recognise sounds under real-world noisy conditions, called robust sound event classification, is an especially challenging task that has attracted recent research attention. Another aspect of real-word conditions is the classification of continuous, occluded or overlapping sounds, rather than classification of short isolated sound recordings. This paper addresses the classification of noise-corrupted, occluded, overlapped, continuous sound recordings. It first proposes a standard evaluation task for such sounds based upon a common existing method for evaluating isolated sound classification. It then benchmarks several high performing isolated sound classifiers to operate with continuous sound data by incorporating an energy-based event detection front end. Results are reported for each tested system using the new task, to provide the first analysis of their performance for continuous sound event detection. In addition it proposes and evaluates a novel Bayesian-inspired front end for the segmentation and detection of continuous sound recordings prior to classification.

Classification of visible and infrared hyperspectral images based on image segmentation and edge-preserving filtering

NASA Astrophysics Data System (ADS)

Cui, Binge; Ma, Xiudan; Xie, Xiaoyun; Ren, Guangbo; Ma, Yi

2017-03-01

The classification of hyperspectral images with a few labeled samples is a major challenge which is difficult to meet unless some spatial characteristics can be exploited. In this study, we proposed a novel spectral-spatial hyperspectral image classification method that exploited spatial autocorrelation of hyperspectral images. First, image segmentation is performed on the hyperspectral image to assign each pixel to a homogeneous region. Second, the visible and infrared bands of hyperspectral image are partitioned into multiple subsets of adjacent bands, and each subset is merged into one band. Recursive edge-preserving filtering is performed on each merged band which utilizes the spectral information of neighborhood pixels. Third, the resulting spectral and spatial feature band set is classified using the SVM classifier. Finally, bilateral filtering is performed to remove "salt-and-pepper" noise in the classification result. To preserve the spatial structure of hyperspectral image, edge-preserving filtering is applied independently before and after the classification process. Experimental results on different hyperspectral images prove that the proposed spectral-spatial classification approach is robust and offers more classification accuracy than state-of-the-art methods when the number of labeled samples is small.

Generalization error analysis: deep convolutional neural network in mammography

NASA Astrophysics Data System (ADS)

Richter, Caleb D.; Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir; Cha, Kenny

2018-02-01

We conducted a study to gain understanding of the generalizability of deep convolutional neural networks (DCNNs) given their inherent capability to memorize data. We examined empirically a specific DCNN trained for classification of masses on mammograms. Using a data set of 2,454 lesions from 2,242 mammographic views, a DCNN was trained to classify masses into malignant and benign classes using transfer learning from ImageNet LSVRC-2010. We performed experiments with varying amounts of label corruption and types of pixel randomization to analyze the generalization error for the DCNN. Performance was evaluated using the area under the receiver operating characteristic curve (AUC) with an N-fold cross validation. Comparisons were made between the convergence times, the inference AUCs for both the training set and the test set of the original image patches without corruption, and the root-mean-squared difference (RMSD) in the layer weights of the DCNN trained with different amounts and methods of corruption. Our experiments observed trends which revealed that the DCNN overfitted by memorizing corrupted data. More importantly, this study improved our understanding of DCNN weight updates when learning new patterns or new labels. Although we used a specific classification task with the ImageNet as example, similar methods may be useful for analysis of the DCNN learning processes, especially those that employ transfer learning for medical image analysis where sample size is limited and overfitting risk is high.

Wearable-Sensor-Based Classification Models of Faller Status in Older Adults.

PubMed

Howcroft, Jennifer; Lemaire, Edward D; Kofman, Jonathan

2016-01-01

Wearable sensors have potential for quantitative, gait-based, point-of-care fall risk assessment that can be easily and quickly implemented in clinical-care and older-adult living environments. This investigation generated models for wearable-sensor based fall-risk classification in older adults and identified the optimal sensor type, location, combination, and modelling method; for walking with and without a cognitive load task. A convenience sample of 100 older individuals (75.5 ± 6.7 years; 76 non-fallers, 24 fallers based on 6 month retrospective fall occurrence) walked 7.62 m under single-task and dual-task conditions while wearing pressure-sensing insoles and tri-axial accelerometers at the head, pelvis, and left and right shanks. Participants also completed the Activities-specific Balance Confidence scale, Community Health Activities Model Program for Seniors questionnaire, six minute walk test, and ranked their fear of falling. Fall risk classification models were assessed for all sensor combinations and three model types: multi-layer perceptron neural network, naïve Bayesian, and support vector machine. The best performing model was a multi-layer perceptron neural network with input parameters from pressure-sensing insoles and head, pelvis, and left shank accelerometers (accuracy = 84%, F1 score = 0.600, MCC score = 0.521). Head sensor-based models had the best performance of the single-sensor models for single-task gait assessment. Single-task gait assessment models outperformed models based on dual-task walking or clinical assessment data. Support vector machines and neural networks were the best modelling technique for fall risk classification. Fall risk classification models developed for point-of-care environments should be developed using support vector machines and neural networks, with a multi-sensor single-task gait assessment.

Land Cover Analysis by Using Pixel-Based and Object-Based Image Classification Method in Bogor

NASA Astrophysics Data System (ADS)

Amalisana, Birohmatin; Rokhmatullah; Hernina, Revi

2017-12-01

The advantage of image classification is to provide earth’s surface information like landcover and time-series changes. Nowadays, pixel-based image classification technique is commonly performed with variety of algorithm such as minimum distance, parallelepiped, maximum likelihood, mahalanobis distance. On the other hand, landcover classification can also be acquired by using object-based image classification technique. In addition, object-based classification uses image segmentation from parameter such as scale, form, colour, smoothness and compactness. This research is aimed to compare the result of landcover classification and its change detection between parallelepiped pixel-based and object-based classification method. Location of this research is Bogor with 20 years range of observation from 1996 until 2016. This region is famous as urban areas which continuously change due to its rapid development, so that time-series landcover information of this region will be interesting.

A new pivoting and iterative text detection algorithm for biomedical images.

PubMed

Xu, Songhua; Krauthammer, Michael

2010-12-01

There is interest to expand the reach of literature mining to include the analysis of biomedical images, which often contain a paper's key findings. Examples include recent studies that use Optical Character Recognition (OCR) to extract image text, which is used to boost biomedical image retrieval and classification. Such studies rely on the robust identification of text elements in biomedical images, which is a non-trivial task. In this work, we introduce a new text detection algorithm for biomedical images based on iterative projection histograms. We study the effectiveness of our algorithm by evaluating the performance on a set of manually labeled random biomedical images, and compare the performance against other state-of-the-art text detection algorithms. We demonstrate that our projection histogram-based text detection approach is well suited for text detection in biomedical images, and that the iterative application of the algorithm boosts performance to an F score of .60. We provide a C++ implementation of our algorithm freely available for academic use. Copyright © 2010 Elsevier Inc. All rights reserved.

Disentangling the brain networks supporting affective speech comprehension.

PubMed

Hervé, Pierre-Yves; Razafimandimby, Annick; Vigneau, Mathieu; Mazoyer, Bernard; Tzourio-Mazoyer, Nathalie

2012-07-16

Areas involved in social cognition, such as the medial prefrontal cortex (mPFC) and the left temporo-parietal junction (TPJ) appear to be active during the classification of sentences according to emotional criteria (happy, angry or sad, [Beaucousin et al., 2007]). These two regions are frequently co-activated in studies about theory of mind (ToM). To confirm that these regions constitute a coherent network during affective speech comprehension, new event-related functional magnetic resonance imaging data were acquired, using the emotional and grammatical-person sentence classification tasks on a larger sample of 51 participants. The comparison of the emotional and grammatical tasks confirmed the previous findings. Functional connectivity analyses established a clear demarcation between a "Medial" network, including the mPFC and TPJ regions, and a bilateral "Language" network, which gathered inferior frontal and temporal areas. These findings suggest that emotional speech comprehension results from interactions between language, ToM and emotion processing networks. The language network, active during both tasks, would be involved in the extraction of lexical and prosodic emotional cues, while the medial network, active only during the emotional task, would drive the making of inferences about the sentences' emotional content, based on their meanings. The left and right amygdalae displayed a stronger response during the emotional condition, but were seldom correlated with the other regions, and thus formed a third entity. Finally, distinct regions belonging to the Language and Medial networks were found in the left angular gyrus, where these two systems could interface. Copyright © 2012 Elsevier Inc. All rights reserved.

A Natural Language Processing-based Model to Automate MRI Brain Protocol Selection and Prioritization.

PubMed

Brown, Andrew D; Marotta, Thomas R

2017-02-01

Incorrect imaging protocol selection can contribute to increased healthcare cost and waste. To help healthcare providers improve the quality and safety of medical imaging services, we developed and evaluated three natural language processing (NLP) models to determine whether NLP techniques could be employed to aid in clinical decision support for protocoling and prioritization of magnetic resonance imaging (MRI) brain examinations. To test the feasibility of using an NLP model to support clinical decision making for MRI brain examinations, we designed three different medical imaging prediction tasks, each with a unique outcome: selecting an examination protocol, evaluating the need for contrast administration, and determining priority. We created three models for each prediction task, each using a different classification algorithm-random forest, support vector machine, or k-nearest neighbor-to predict outcomes based on the narrative clinical indications and demographic data associated with 13,982 MRI brain examinations performed from January 1, 2013 to June 30, 2015. Test datasets were used to calculate the accuracy, sensitivity and specificity, predictive values, and the area under the curve. Our optimal results show an accuracy of 82.9%, 83.0%, and 88.2% for the protocol selection, contrast administration, and prioritization tasks, respectively, demonstrating that predictive algorithms can be used to aid in clinical decision support for examination protocoling. NLP models developed from the narrative clinical information provided by referring clinicians and demographic data are feasible methods to predict the protocol and priority of MRI brain examinations. Copyright © 2017 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.

Driver behavior profiling: An investigation with different smartphone sensors and machine learning

PubMed Central

Ferreira, Jair; Carvalho, Eduardo; Ferreira, Bruno V.; de Souza, Cleidson; Suhara, Yoshihiko; Pentland, Alex

2017-01-01

Driver behavior impacts traffic safety, fuel/energy consumption and gas emissions. Driver behavior profiling tries to understand and positively impact driver behavior. Usually driver behavior profiling tasks involve automated collection of driving data and application of computer models to generate a classification that characterizes the driver aggressiveness profile. Different sensors and classification methods have been employed in this task, however, low-cost solutions and high performance are still research targets. This paper presents an investigation with different Android smartphone sensors, and classification algorithms in order to assess which sensor/method assembly enables classification with higher performance. The results show that specific combinations of sensors and intelligent methods allow classification performance improvement. PMID:28394925

Localizing tuberculosis in chest radiographs with deep learning

NASA Astrophysics Data System (ADS)

Xue, Zhiyun; Jaeger, Stefan; Antani, Sameer; Long, L. Rodney; Karargyris, Alexandros; Siegelman, Jenifer; Folio, Les R.; Thoma, George R.

2018-03-01

Chest radiography (CXR) has been used as an effective tool for screening tuberculosis (TB). Because of the lack of radiological expertise in resource-constrained regions, automatic analysis of CXR is appealing as a "first reader". In addition to screening the CXR for disease, it is critical to highlight locations of the disease in abnormal CXRs. In this paper, we focus on the task of locating TB in CXRs which is more challenging due to the intrinsic difficulty of locating the abnormality. The method is based on applying a convolutional neural network (CNN) to classify the superpixels generated from the lung area. Specifically, it consists of four major components: lung ROI extraction, superpixel segmentation, multi-scale patch generation/labeling, and patch classification. The TB regions are located by identifying those superpixels whose corresponding patches are classified as abnormal by the CNN. The method is tested on a publicly available TB CXR dataset which contains 336 TB images showing various manifestations of TB. The TB regions in the images were marked by radiologists. To evaluate the method, the images are split into training, validation, and test sets with all the manifestations being represented in each set. The performance is evaluated at both the patch level and image level. The classification accuracy on the patch test set is 72.8% and the average Dice index for the test images is 0.67. The factors that may contribute to misclassification are discussed and directions for future work are addressed.

Phase contrast imaging X-ray computed tomography: quantitative characterization of human patellar cartilage matrix with topological and geometrical features

NASA Astrophysics Data System (ADS)

Nagarajan, Mahesh B.; Coan, Paola; Huber, Markus B.; Diemoz, Paul C.; Wismüller, Axel

2014-03-01

Current assessment of cartilage is primarily based on identification of indirect markers such as joint space narrowing and increased subchondral bone density on x-ray images. In this context, phase contrast CT imaging (PCI-CT) has recently emerged as a novel imaging technique that allows a direct examination of chondrocyte patterns and their correlation to osteoarthritis through visualization of cartilage soft tissue. This study investigates the use of topological and geometrical approaches for characterizing chondrocyte patterns in the radial zone of the knee cartilage matrix in the presence and absence of osteoarthritic damage. For this purpose, topological features derived from Minkowski Functionals and geometric features derived from the Scaling Index Method (SIM) were extracted from 842 regions of interest (ROI) annotated on PCI-CT images of healthy and osteoarthritic specimens of human patellar cartilage. The extracted features were then used in a machine learning task involving support vector regression to classify ROIs as healthy or osteoarthritic. Classification performance was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC). The best classification performance was observed with high-dimensional geometrical feature vectors derived from SIM (0.95 ± 0.06) which outperformed all Minkowski Functionals (p < 0.001). These results suggest that such quantitative analysis of chondrocyte patterns in human patellar cartilage matrix involving SIM-derived geometrical features can distinguish between healthy and osteoarthritic tissue with high accuracy.

Detection of white matter lesion regions in MRI using SLIC0 and convolutional neural network.

PubMed

Diniz, Pedro Henrique Bandeira; Valente, Thales Levi Azevedo; Diniz, João Otávio Bandeira; Silva, Aristófanes Corrêa; Gattass, Marcelo; Ventura, Nina; Muniz, Bernardo Carvalho; Gasparetto, Emerson Leandro

2018-04-19

White matter lesions are non-static brain lesions that have a prevalence rate up to 98% in the elderly population. Because they may be associated with several brain diseases, it is important that they are detected as soon as possible. Magnetic Resonance Imaging (MRI) provides three-dimensional data with the possibility to detect and emphasize contrast differences in soft tissues, providing rich information about the human soft tissue anatomy. However, the amount of data provided for these images is far too much for manual analysis/interpretation, representing a difficult and time-consuming task for specialists. This work presents a computational methodology capable of detecting regions of white matter lesions of the brain in MRI of FLAIR modality. The techniques highlighted in this methodology are SLIC0 clustering for candidate segmentation and convolutional neural networks for candidate classification. The methodology proposed here consists of four steps: (1) images acquisition, (2) images preprocessing, (3) candidates segmentation and (4) candidates classification. The methodology was applied on 91 magnetic resonance images provided by DASA, and achieved an accuracy of 98.73%, specificity of 98.77% and sensitivity of 78.79% with 0.005 of false positives, without any false positives reduction technique, in detection of white matter lesion regions. It is demonstrated the feasibility of the analysis of brain MRI using SLIC0 and convolutional neural network techniques to achieve success in detection of white matter lesions regions. Copyright © 2018. Published by Elsevier B.V.

Selecting relevant 3D image features of margin sharpness and texture for lung nodule retrieval.

PubMed

Ferreira, José Raniery; de Azevedo-Marques, Paulo Mazzoncini; Oliveira, Marcelo Costa

2017-03-01

Lung cancer is the leading cause of cancer-related deaths in the world. Its diagnosis is a challenge task to specialists due to several aspects on the classification of lung nodules. Therefore, it is important to integrate content-based image retrieval methods on the lung nodule classification process, since they are capable of retrieving similar cases from databases that were previously diagnosed. However, this mechanism depends on extracting relevant image features in order to obtain high efficiency. The goal of this paper is to perform the selection of 3D image features of margin sharpness and texture that can be relevant on the retrieval of similar cancerous and benign lung nodules. A total of 48 3D image attributes were extracted from the nodule volume. Border sharpness features were extracted from perpendicular lines drawn over the lesion boundary. Second-order texture features were extracted from a cooccurrence matrix. Relevant features were selected by a correlation-based method and a statistical significance analysis. Retrieval performance was assessed according to the nodule's potential malignancy on the 10 most similar cases and by the parameters of precision and recall. Statistical significant features reduced retrieval performance. Correlation-based method selected 2 margin sharpness attributes and 6 texture attributes and obtained higher precision compared to all 48 extracted features on similar nodule retrieval. Feature space dimensionality reduction of 83 % obtained higher retrieval performance and presented to be a computationaly low cost method of retrieving similar nodules for the diagnosis of lung cancer.

Fractal analysis of seafloor textures for target detection in synthetic aperture sonar imagery

NASA Astrophysics Data System (ADS)

Nabelek, T.; Keller, J.; Galusha, A.; Zare, A.

2018-04-01

Fractal analysis of an image is a mathematical approach to generate surface related features from an image or image tile that can be applied to image segmentation and to object recognition. In undersea target countermeasures, the targets of interest can appear as anomalies in a variety of contexts, visually different textures on the seafloor. In this paper, we evaluate the use of fractal dimension as a primary feature and related characteristics as secondary features to be extracted from synthetic aperture sonar (SAS) imagery for the purpose of target detection. We develop three separate methods for computing fractal dimension. Tiles with targets are compared to others from the same background textures without targets. The different fractal dimension feature methods are tested with respect to how well they can be used to detect targets vs. false alarms within the same contexts. These features are evaluated for utility using a set of image tiles extracted from a SAS data set generated by the U.S. Navy in conjunction with the Office of Naval Research. We find that all three methods perform well in the classification task, with a fractional Brownian motion model performing the best among the individual methods. We also find that the secondary features are just as useful, if not more so, in classifying false alarms vs. targets. The best classification accuracy overall, in our experimentation, is found when the features from all three methods are combined into a single feature vector.

Automated radial basis function neural network based image classification system for diabetic retinopathy detection in retinal images

NASA Astrophysics Data System (ADS)

Anitha, J.; Vijila, C. Kezi Selva; Hemanth, D. Jude

2010-02-01

Diabetic retinopathy (DR) is a chronic eye disease for which early detection is highly essential to avoid any fatal results. Image processing of retinal images emerge as a feasible tool for this early diagnosis. Digital image processing techniques involve image classification which is a significant technique to detect the abnormality in the eye. Various automated classification systems have been developed in the recent years but most of them lack high classification accuracy. Artificial neural networks are the widely preferred artificial intelligence technique since it yields superior results in terms of classification accuracy. In this work, Radial Basis function (RBF) neural network based bi-level classification system is proposed to differentiate abnormal DR Images and normal retinal images. The results are analyzed in terms of classification accuracy, sensitivity and specificity. A comparative analysis is performed with the results of the probabilistic classifier namely Bayesian classifier to show the superior nature of neural classifier. Experimental results show promising results for the neural classifier in terms of the performance measures.

Posture recognition associated with lifting of heavy objects using Kinect and Adaboost

NASA Astrophysics Data System (ADS)

Raut, Sayli; Navaneethakrishna, M.; Ramakrishnan, S.

2017-12-01

Lifting of heavy objects is the common task in the industries. Recent statistics from the Bureau of Labour indicate, back injuries account for one of every five injuries in the workplace. Eighty per cent of these injuries occur to the lower back and are associated with manual materials handling tasks. According to the Industrial ergonomic safety manual, Squatting is the correct posture for lifting a heavy object. In this work, an attempt has been made to monitor posture of the workers during squat and stoop using 3D motion capture and machine learning techniques. For this, Microsoft Kinect V2 is used for capturing the depth data. Further, Dynamic Time Warping and Euclidian distance algorithms are used for extraction of features. Ada-boost algorithm is used for classification of stoop and squat. The results show that the 3D image data is large and complex to analyze. The application of nonlinear and linear metrics captures the variation in the lifting pattern. Additionally, the features extracted from this metric resulted in a classification accuracy of 85% and 81% respectively. This framework may be put-upon to alert the workers in the industrial ergonomic environments.

Feature selection and classification of multiparametric medical images using bagging and SVM

NASA Astrophysics Data System (ADS)

Fan, Yong; Resnick, Susan M.; Davatzikos, Christos

2008-03-01

This paper presents a framework for brain classification based on multi-parametric medical images. This method takes advantage of multi-parametric imaging to provide a set of discriminative features for classifier construction by using a regional feature extraction method which takes into account joint correlations among different image parameters; in the experiments herein, MRI and PET images of the brain are used. Support vector machine classifiers are then trained based on the most discriminative features selected from the feature set. To facilitate robust classification and optimal selection of parameters involved in classification, in view of the well-known "curse of dimensionality", base classifiers are constructed in a bagging (bootstrap aggregating) framework for building an ensemble classifier and the classification parameters of these base classifiers are optimized by means of maximizing the area under the ROC (receiver operating characteristic) curve estimated from their prediction performance on left-out samples of bootstrap sampling. This classification system is tested on a sex classification problem, where it yields over 90% classification rates for unseen subjects. The proposed classification method is also compared with other commonly used classification algorithms, with favorable results. These results illustrate that the methods built upon information jointly extracted from multi-parametric images have the potential to perform individual classification with high sensitivity and specificity.

Visualizing histopathologic deep learning classification and anomaly detection using nonlinear feature space dimensionality reduction.

PubMed

Faust, Kevin; Xie, Quin; Han, Dominick; Goyle, Kartikay; Volynskaya, Zoya; Djuric, Ugljesa; Diamandis, Phedias

2018-05-16

There is growing interest in utilizing artificial intelligence, and particularly deep learning, for computer vision in histopathology. While accumulating studies highlight expert-level performance of convolutional neural networks (CNNs) on focused classification tasks, most studies rely on probability distribution scores with empirically defined cutoff values based on post-hoc analysis. More generalizable tools that allow humans to visualize histology-based deep learning inferences and decision making are scarce. Here, we leverage t-distributed Stochastic Neighbor Embedding (t-SNE) to reduce dimensionality and depict how CNNs organize histomorphologic information. Unique to our workflow, we develop a quantitative and transparent approach to visualizing classification decisions prior to softmax compression. By discretizing the relationships between classes on the t-SNE plot, we show we can super-impose randomly sampled regions of test images and use their distribution to render statistically-driven classifications. Therefore, in addition to providing intuitive outputs for human review, this visual approach can carry out automated and objective multi-class classifications similar to more traditional and less-transparent categorical probability distribution scores. Importantly, this novel classification approach is driven by a priori statistically defined cutoffs. It therefore serves as a generalizable classification and anomaly detection tool less reliant on post-hoc tuning. Routine incorporation of this convenient approach for quantitative visualization and error reduction in histopathology aims to accelerate early adoption of CNNs into generalized real-world applications where unanticipated and previously untrained classes are often encountered.

Flightspeed Integral Image Analysis Toolkit

NASA Technical Reports Server (NTRS)

Thompson, David R.

2009-01-01

The Flightspeed Integral Image Analysis Toolkit (FIIAT) is a C library that provides image analysis functions in a single, portable package. It provides basic low-level filtering, texture analysis, and subwindow descriptor for applications dealing with image interpretation and object recognition. Designed with spaceflight in mind, it addresses: Ease of integration (minimal external dependencies) Fast, real-time operation using integer arithmetic where possible (useful for platforms lacking a dedicated floatingpoint processor) Written entirely in C (easily modified) Mostly static memory allocation 8-bit image data The basic goal of the FIIAT library is to compute meaningful numerical descriptors for images or rectangular image regions. These n-vectors can then be used directly for novelty detection or pattern recognition, or as a feature space for higher-level pattern recognition tasks. The library provides routines for leveraging training data to derive descriptors that are most useful for a specific data set. Its runtime algorithms exploit a structure known as the "integral image." This is a caching method that permits fast summation of values within rectangular regions of an image. This integral frame facilitates a wide range of fast image-processing functions. This toolkit has applicability to a wide range of autonomous image analysis tasks in the space-flight domain, including novelty detection, object and scene classification, target detection for autonomous instrument placement, and science analysis of geomorphology. It makes real-time texture and pattern recognition possible for platforms with severe computational restraints. The software provides an order of magnitude speed increase over alternative software libraries currently in use by the research community. FIIAT can commercially support intelligent video cameras used in intelligent surveillance. It is also useful for object recognition by robots or other autonomous vehicles