The ability to differentiate sensor-induced physiological signals between healthy and diseased subjects is useful for developing an e-health system. Patients with neurodegenerative disorders are among those who can benefit from the use of e-health. Entropy methods have been utilized to quantify the complexity of such physiological signals for pattern classification. To date, these methods have been applied individually. In this retrospective study, several entropy methods are examined and used as feature extraction methods for machine learning to classify gait patterns in neurodegenerative diseases. Experimental results show that the combination of entropy methods and standard statistical measures performed much better than the individual measures for physiological pattern differentiation. Several machine learning models were also evaluated for learning on these features. This study also found that the one-dimensional convolutional network model trained with the combined features provided the most favorable results, where the best entropy mea-sures depend on certain values for time delays and embedding dimensions.

The free-living nematode Caenorhabditis elegans is an ideal model for understanding behavior and networks of neurons. Experimental and quantitative analyses of neural circuits and behavior have led to system-level understanding of behavioral genetics and process of transformation from sensory integration in stimulus environments to behavioral outcomes. The ability to differentiate locomotion behavior between wild-type and mutant Caenorhabditis elegans strains allows precise inference on and gaining insights into genetic and environmental influences on behaviors. This paper presents an eigenfeature-enhanced deep-learning method for classifying the dynamics of locomotion behavior of wild-type and mutant Caenorhabditis elegans. Classification results obtained from public benchmark time-series data of eigenworms illustrate the superior performance of the new method over several existing classifiers. The proposed method has potential as a useful artificial-intelligence tool for automated identification of the nematode worm behavioral patterns aiming at elucidating molecular and genetic mechanisms that control the nervous system.

The ability to predict survival in cancer is clinically important because the finding can help patients and physicians make optimal treatment decisions. Artificial intelligence in the context of deep learning has been increasingly realized by the informatics-oriented medical community as a powerful machine-learning technology for cancer research, diagnosis, prediction, and treatment. This paper presents the combination of deep learning, data coding, and probabilistic modeling for predicting five-year survival in a cohort of patients with rectal cancer using images of RhoB expression on biopsies. Using about one-third of the patients' data for testing, the proposed approach achieved 90% prediction accuracy, which is much higher than the direct use of the best pretrained convolutional neural network (70%) and the best coupling of a pretrained model and support vector machines (70%).

A literature survey shows that the number of malware attacks is gradually growing over the years due to the growing trend of Internet of Medical Things (IoMT) devices. To detect and classify malware attacks, automated malware detection and classification is an essential subsystem in healthcare cyber-physical systems. This work proposes an attention-based multidimensional deep learning (DL) approach for a cross-architecture IoMT malware detection and classification system based on byte sequences extracted from Executable and Linkable Format (ELF; formerly named Extensible Linking Format) files. The DL approach automates the feature design and extraction process from unstructured byte sequences. In addition, the proposed approach facilitates the detection of the central processing unit (CPU) architecture of the ELF file. A detailed experimental analysis and its evaluation are shown on the IoMT cross-architecture benchmark dataset. In all the experiments, the proposed method showed better performance compared with those obtained from several existing methods with an accuracy of 95% for IoMT malware detection, 94% for IoMT malware classification, and 95% for CPU architectures classification. The proposed method also suggests a similar performance with an accuracy of 94% on the Microsoft malware dataset. Experimental results on two malware datasets indicate that the proposed method is robust and generalizable in cross-architecture IoMT malware detection, classification, and CPU architectures classification in healthcare cyber-physical systems.

The ability to detect gait in Parkinson’s disease (PD) is useful for the diagnosis of PD and monitoring its progress. For the first time, based on an artificial intelligence-based notion of graphical temporal recurrence modeling of nonlinear systems, this paper presents a visual concurrent analysis of gait dynamics for detecting visual changes in patterns among young, old, and PD subjects. More specifically, the analysis is based on tensor decomposition of fuzzy recurrence plots of wearable sensor-induced stride signals. After validating the proposed approach using time series of known systems, results obtained for the analysis of gait patterns among the three cohorts suggest the combined methods as a potential computerized health data analysis tool for early recognition, assessment, and management of PD in either clinical or home environments.

Rectal or colorectal cancer is one of the leading causes of cancer-related death. With the advancement in surgical techniques, the survival rate has been improved. Predicting the survival rate is an important factor for enabling optimal treatments to prolong rectal-cancer patients' lives. Methods of artificial intelligence and machine learning have been applied for assisting physicians in cancer research. In this study, we investigated the use of pretrained convolutional neural networks and support vector machines for predicting the survival rate of a cohort of rectal-cancer patients using metastatic immunohistochemistry samples staining for protein RhoB. The combination of convolutional neural networks and support vector machines achieved better classification results than using individual pretrained deep networks in most cases, and where manual pathological analysis is encountered with great difficulty. In particular, the combination of ResNet-101 and SVM produced an average accuracy of 86% for non-radiotherapy, and Inception-v3 and SVM resulted in an average accuracy of 85% for radiotherapy.

Abstract With the popularity of the internet and smartphones, malware on smartphones has increased dramatically. In addition, the ubiquity and openness of the Android operating system have made it a lucrative platform for cybercriminals to develop malware. Traditional malware detection techniques require a lot of time and manual effort to classify malware accurately. Recently, deep learning (DL) based malware detection and classification techniques have been developed to solve this issue. This article proposes a DL‐based two‐stage framework that detects Android malware and classifies its variants using image‐based malware representations of the Android DEX files. The framework uses the EfficientNetB0 convolutional neural network (CNN) to extracts relevant features from the malware color images. The extracted features are then passed through a global average pooling layer and fed into a stacking classifier. The stacking classifier employs linear support vector machine (SVM) and random forest (RF) algorithms as base‐level classifiers and logistic regression as the meta‐level classifier. This method obtained an accuracy of 100% in the binary classification of Android malware images and a 92.9% accuracy in 5‐class (Adsware, Adware + Adware, Clicker + Trojan, Spyware, and Benign) classification, and an 88.6% accuracy in 4‐class (Adsware, Adware + Adware, Clicker + Trojan, and Spyware) classification. We compared our method with 26 state‐of‐the‐art pretrained CNN models (including the original EfficientNetB0) and large‐scale learning classifiers such as EfficientNetB0‐SVM and EfficientNetB0‐RF. The proposed framework outperformed the compared methods in all performance metrics. Experiments also demonstrate that substituting the softmax layer of CNNs with a large‐scale learning classifier or stacking classifier results in an enhanced performance over the original network.

Abstract Cassava is a rich source of carbohydrates, and it is vulnerable to virus diseases. Literature survey shows that the image recognition and integrated deep learning approach is successfully employed for Cassava leaf disease classification. Mostly, transfer learning based on a convolutional neural network (CNN) models were successfully applied for Cassava leaf disease classification. However, existing approaches are not effective in identifying the tiny portion of the disease in the overall leaf area. Identifying and focussing on regions affected by the disease is vital to achieving a good classification accuracy. An attention‐based approach is integrated into pretrained CNN‐based EfficientNet models to locate and identify the tiny infected regions in Cassava leaf. Penultimate layer features of attention‐based EfficientNet models such as A_EfficientNetB4, A_EfficientNetB5, and A_EfficientNetB6 were extracted. Next, the dimensionality of the extracted features was reduced using kernel principal component analysis. The reduced features were fused and passed into a stacked ensemble meta‐classifier for Cassava leaf disease classification. A stacked ensemble meta‐classifier is a two‐stage approach in which the first stage employs random forest and support vector machine (SVM) for prediction followed by logistic regression for classification. Detailed investigation and analysis of the proposed method, attention, and non‐attention‐based approaches with CNN pretrained models were tested using a publicly available benchmark dataset of Cassava leaf disease images. The proposed method achieved better performances in all experiments than several existing methods as well as various attention and non‐attention‐based CNN pretrained models. The proposed approach can be used as a deployable tool for Cassava leaf disease classification in agricultural field.

Abstract Literature survey shows that convolutional neural network (CNN)‐based pretrained models have been successfully employed to diagnose and detect childhood pneumonia using chest X‐rays (CXR). However, most of the existing methods are prone to imbalance problems, which become even more significant in medical image classification for example most importantly childhood pneumonia classification using CXR. This is due to the fact that some classes in childhood pneumonia have a very little support in the training dataset. Additionally, though the existing methods have reported better performances for training and testing, in most of the test cases the existing models will not be effective on variants of the childhood pneumonia CXR images or CXR samples from a new pediatric patient. In addition, the models may be effective in detecting latent stage pediatric pneumonia but not show better performances for CXR samples from pediatric patients who are early stage, sick but not pneumonia, sick with other lung diseases, and so on. Generalization is an important term to be considered while designing a pneumonia classifier that can perform well on completely unseen pneumonia CXR datasets. This article presents a cost‐sensitive large‐scale learning with stacked ensemble meta‐classifier and transfer learning‐based deep feature fusion approach for pediatric pneumonia classification using CXR. With the aim to identify the importance among the classes of pneumonia, the larger cost items are introduced based on the class‐imbalance degree during the backpropogation learning methodology in transfer learning models such as Xception, InceptionResNetV2, DenseNet201, and NASNetMobile. Next, the features from the penultimate layer (global average pooling) of Xception, InceptionResNetV2, DenseNet201, and NASNetMobile were extracted and dimensionality of the extracted features were reduced using kernel principal component analysis (KPCA). The reduced features were fused together and passed into a stacked ensemble meta‐classifier for classifying the CXR into either pneumonia or normal. A stacked ensemble meta‐classifier is a two stage approach in which the first stage employs random forest and support vector machine (SVM) for prediction and followed by logistic regression for classification. Experiments of the proposed model were done on publicly available benchmark pediatric pneumonia classification CXR dataset. In addition, the experiments for existing methods as well as various cost‐insensitive models were conducted. In all the experiments, the proposed method has achieved better performances compared to the existing methods as well as various cost‐insensitive models. In particular, the proposed method showed 6% improvement in precision, 10% improvement in recall, 9% improvement in F1 score with less misclassification costs (0.0321) and accuracy (96.8%). Most importantly, the proposed method is insensitive to the imbalance data and more effective to handle variants of the childhood pneumonia CXR images. Thus, the proposed approach can be used as a tool for point‐of‐care diagnosis by healthcare professionals.

This paper presents the development of a non-invasive image-analysis method for improving the diagnostic accuracy of mediastinal lymph node metastasis in patients with lung cancer. The approach adopts pretrained deep learning models incorporated with the geostatistical simulation of texture of benign and metastatic lung lymph nodes in computed tomography (CT) images for classification. Using 271 CT samples of mediastinal lymph nodes collected from 148 patients with lung cancer, deep-learning models coupled with the stochastic simulated data augmentation provide the best classification results. The simulation of texture in these medical images was able to discover rich radiomic features to ascertain subtle difference between benign and metastatic lymph nodes and enhance the performance of the deep-learning models for complex pattern classification. The proposed approach is very promising to be utilized as a computerized tool for medical image analysis.

By applying deep learning, we were able to compare p73 protein expression patterns of different tissue types including normal mucosa, primary tumor and lymph node metastasis in rectal cancer patients using immunohistochemical slides. The pair-wise pattern comparisons were automatedly carried out by considering color, edge, blobs, and other morphological information in the images. We discovered that when the pattern dissimilarity between primary tumor and lymph node metastasis is relatively low among other tissue pairs (primary tumor and distant normal, biopsy and distant normal, biopsy and primary tumor, biopsy and primary tumor, lymph node metastasis and distant normal, lymph node metastasis and biopsy), there was an implication of short-time survival. This original result suggests a novel application of advanced artificial intelligence in machine learning for clinical finding in rectal cancer and encourages relevant study of multiple biomarker expressions in cancer patients.

Photoplethysmogram (PPG) signals obtained from wearable sensors have been utilized for monitoring health conditions in both clinical and non-clinical environments, mostly concerning with heart-rate events. This paper shows the potential use of short-time PPG signals for differentiating patients with Parkinson's disease (PD) from healthy control (HC) subjects with nonlinear dynamics analysis. Multiscale entropy, time-shift multiscale entropy, and fuzzy recurrence plots were applied for extracting features from PPG signals of PD patients and HC subjects. Least-square support vector machine based cross-validations of the features extracted from the three nonlinear dynamics analysis methods achieve high classification rates, where those obtained from fuzzy recurrence plots are the highest.

The importance of developing effective strategies for investigating mediastinal lymph-node metastases in non-small cell lung cancers is increasingly emphasized. It is because the precise detection of this metastatic disease is critical for optimal surgical intervention and treatment for patients with lung cancer. Existing medical image analysis is of limited power for mediastinal lymph-node staging on computed tomography (CT). Motivated by the radiomics hypothesis, this paper explored deep-learning, texture features and their combinations to ascertain subtle difference between malignant and benign mediastinal lymph nodes on CT. The radiomics-based results are found to be promising for differentiating malignant from benign mediastinal lymph nodes of patients with lung cancer.

Most texture analysis techniques require training data to perform classification or retrieval of images. In many practical situations, the amount of data representing different texture classes can be too limited to satisfy the training of a reliable classifier. Therefore, finding an effective feature of texture is very useful to cope with a variety of applications. This paper presents the extension of the two-point variogram to multiple-point variogram of images for texture feature extraction, which is also robust to noise and computationally economic. The matching of the variogram functions for pattern classification can be enhanced with the use of a spectral distortion measure without the requirement of training data. Experimental results and comparison with other methods, which require training data, suggest the usefulness of the proposed approach.

This paper presents a novel method for performing guided image filtering using local polynomial approximation (LPA) with range guidance. In our method, the LPA is introduced into a multipoint framework for reliable model regression and better preservation on image spatial variation which usually contains the essential information in the input image. In addition, we develop a weighting scheme which has the spatial flexibility during the filtering process. All components in our method are efficiently implemented and a constant computation complexity is achieved. Compared with conventional filtering methods, our method provides clearer boundaries and performs especially better in recovering spatial variation from noisy images. We conduct a number of experiments for different applications: depth image upsampling, joint image denoising, details enhancement, and image abstraction. Both quantitative and qualitative comparisons demonstrate that our method outperforms state-of-the-art methods.

Semi-variogram estimators and distortion measures of signal spectra are utilized in this paper for image texture retrieval. On the use of the complete Brodatz database, most high retrieval rates are reportedly based on multiple features and the combinations of multiple algorithms, while the classification using single features is still a challenge to the retrieval of diverse texture images. The semi-variogram, which is theoretically sound and the cornerstone of spatial statistics, has the characteristics shared between true randomness and complete determinism and, therefore, can be used as a useful tool for both the structural and statistical analysis of texture images. Meanwhile, spectral distortion measures derived from the theory of linear predictive coding provide a rigorously mathematical model for signal-based similarity matching and have been proven useful for many practical pattern classification systems. Experimental results obtained from testing the proposed approach using the complete Brodatz database, and the the University of Illinois at Urbana-Champaign texture database suggests the effectiveness of the proposed approach as a single-feature-based dissimilarity measure for real-time texture retrieval.

Filtering additive white Gaussian noise in images using the best linear unbiased estimator (BLUE) is technically sound in a sense that it is an optimal average filter derived from the statistical estimation theory. The BLUE filter mask has the theoretical advantage in that its shape and its size are formulated in terms of the image signals and associated noise components. However, like many other noise filtering problems, prior knowledge about the additive noise needs to be available, which is often obtained using training data. This paper presents the sequential Gaussian simulation in geostatistics for measuring signal and noise variances in images without the need of training data for the BLUE filter implementation. The simulated signal variance and the BLUE average can be further used as parameters of the adaptive Wiener filter for image restoration.

This paper presents a novel guided image filtering method using multipoint local polynomial approximation (LPA) with range guidance. In our method, the LPA is extended from a pointwise model into a multipoint model for reliable filtering and better preserving image spatial variation which usually contains the essential information in the input image. In addition, we develop a scheme with constant computational complexity (invariant to the size of filtering kernel) for generating a spatial adaptive support region around a point. By using the hybrid of the local polynomial model and color/intensity based range guidance, the proposed method not only preserves edges but also does a much better job in preserving spatial variation than existing popular filtering methods. Our method proves to be effective in a number of applications: depth image upsampling, joint image denoising, details enhancement, and image abstraction. Experimental results show that our method produces better results than state-of-the-art methods and it is also computationally efficient.

Intestinal abnormalities and ischemia are medical conditions in which inflammation and injury of the intestine are caused by inadequate blood supply. Developments of computerized systems for the automated identification of these types of complex gastrointestinal disorders are rarely reported. In this paper, we introduce a mapping model of spatial uncertainty in computed tomography images for feature extraction, which can be effectively applied for diagnostic detection. Experimental results obtained from the analysis of clinical data suggest the usefulness of the proposed uncertainty mapping model.

With the massive production of genomic and proteomic data, the number of available biological sequences in databases has reached a level that is not feasible anymore for exact alignments even when just a fraction of all sequences is used. To overcome this inevitable time complexity, ultrafast alignment-free methods are studied. Within the past two decades, a broad variety of nonalignment methods have been proposed including dissimilarity measures on classical representations of sequences like k-words or Markov models. Furthermore, articles were published that describe distance measures on alternative representations such as compression complexity, spectral time series or chaos game representation. However, alignments are still the standard method for real world applications in biological sequence analysis, and the time efficient alignment-free approaches are usually applied in cases when the accustomed algorithms turn out to fail or be too inconvenient.

Texture analysis is a major research topic in intelligent image processing. Its useful applications, including object detection and classification, to various fields of engineering, science, medicine, biology, and creative arts have been increasingly reported. Texture is a fundamental aspect of human vision and perception by which different types of objects can be distinguished through their appearance, ranging from distinctive to subtle roughness. Given tremendous efforts in terms of both theoretical developments and applications, texture analysis still remains a challenging area of research in image analysis and pattern recognition. This paper presents a novel and practical image texture analysis method using the fundamentals of geostatistics and the concept of entropy in information theory. Experimental results on medical and document image data have shown the superior performance of the proposed approach over its related texture analysis methods.

One of the most challenging problems in the study and interpretation of complex biological data is the quantification of the dynamical mechanisms and complexity of the distributions of cells and their organelles in time and space. Such quantification has important implications for computer modeling and simulation of diseases. This paper presents an original investigation into the potential chaos of the dynamics underlying the endoplasmic-reticulum network recorded in time-lapse microscopy images.

Dementia is a most common neurodegenerative disorder. Previous researches have attempted to relate the impairment of cognition to volumes or thickness changes of the cortical regions, with relatively few studies investigating other features such as cortical surface anatomy. In the present study, we report for the first time to use sample entropy (SampEn) to assess the complexity of cortical surface structure in early stage of dementia compared to healthy controls. Whole brain structural T1-weighted MRI scans were collected from 192 subjects including patients with very mild and mild dementia (age = 77 ± 7, male/female = 41/55, CDR = 0.5 or 1, n = 96), and healthy subjects (aged = 76 ± 9, male/female= 25/70, CDR = 0, n = 96). SampEn was applied to each transection and averaged. Comparisons were made between control and dementia for the whole brain as well as for each sub-section of all layers. Results show an overall larger SampEn in demented group compared with non-demented group(p <; 0.05) which indicate an increase of structural irregularity of cortical surface in dementia. Our findings offer a novel approach for studying cortical atrophy patterns and can potentially be used to develop novel biomarkers of dementia.

Information of cellular organelle location and morphology is essential for cancer simulation. In order to obtain such information, the segmentation of the organelles from electronic microscopy intracellular image is crucial. In this paper, we focus on the segmentation of mitochondria organelle which is one of the most important organelles tightly related to the form of cancer. A simple three-stage strategy for mitochondrial segmentation based on exclusive and morphology properties and Gabor filter is proposed. Experimental results on focused ion beam (FIB) and scanning electron microscope (SEM) images have shown the effectiveness of proposed method.

It has been hypothesized that by examining the changes of intensity in medical images, we can extract some property that characterizes the pathology of disease process. This paper presents the conceptual framework of the regularity dimension of images which can be utilized in computational neuroscience to obtain objective information to support clinical decision making. As an example model, the proposed approach is applied for studying pattern similarity of white matter hyperintensities of the brain on magnetic resonance imaging.

BACKGROUND: Automated, image based high-content screening is a fundamental tool for discovery in biological science. Modern robotic fluorescence microscopes are able to capture thousands of images from massively parallel experiments such as RNA interference (RNAi) or small-molecule screens. As such, efficient computational methods are required for automatic cellular phenotype identification capable of dealing with large image data sets. In this paper we investigated an efficient method for the extraction of quantitative features from images by combining second order statistics, or Haralick features, with curvelet transform. A random subspace based classifier ensemble with multiple layer perceptron (MLP) as the base classifier was then exploited for classification. Haralick features estimate image properties related to second-order statistics based on the grey level co-occurrence matrix (GLCM), which has been extensively used for various image processing applications. The curvelet transform has a more sparse representation of the image than wavelet, thus offering a description with higher time frequency resolution and high degree of directionality and anisotropy, which is particularly appropriate for many images rich with edges and curves. A combined feature description from Haralick feature and curvelet transform can further increase the accuracy of classification by taking their complementary information. We then investigate the applicability of the random subspace (RS) ensemble method for phenotype classification based on microscopy images. A base classifier is trained with a RS sampled subset of the original feature set and the ensemble assigns a class label by majority voting. RESULTS: Experimental results on the phenotype recognition from three benchmarking image sets including HeLa, CHO and RNAi show the effectiveness of the proposed approach. The combined feature is better than any individual one in the classification accuracy. The ensemble model produces better classification performance compared to the component neural networks trained. For the three images sets HeLa, CHO and RNAi, the Random Subspace Ensembles offers the classification rates 91.20%, 98.86% and 91.03% respectively, which compares sharply with the published result 84%, 93% and 82% from a multi-purpose image classifier WND-CHARM which applied wavelet transforms and other feature extraction methods. We investigated the problem of estimation of ensemble parameters and found that satisfactory performance improvement could be brought by a relative medium dimensionality of feature subsets and small ensemble size. CONCLUSIONS: The characteristics of curvelet transform of being multiscale and multidirectional suit the description of microscopy images very well. It is empirically demonstrated that the curvelet-based feature is clearly preferred to wavelet-based feature for bioimage descriptions. The random subspace ensemble of MLPs is much better than a number of commonly applied multi-class classifiers in the investigated application of phenotype recognition.

Detection of white matter changes of the brain using magnetic resonance imaging (MRI) has increasingly been an active and challenging research area in computational neuroscience. There have rarely been any single image analysis methods that can effectively address the issue of automated quantification of neuroimages, which are subject to different interests of various medical hypotheses. This paper presents new image segmentation models for automated detection of white matter changes of the brain in an elderly population. The methods are based on the computational models of fuzzy clustering, possibilistic clustering, geostatistics, and knowledge combination. Experimental results on MRI data have shown that the proposed image analysis methodology can be applied as a very useful computerized tool for the validation of our particular medical question, where white matter changes of the brain are thought to be the most important social medical evidence.

Motivated by fuzzy clustering incorporating spatial information, we present a spatially constrained fuzzy hyper-prototype clustering algorithm in this paper. This approach uses hyperplanes as cluster centers and adds a spatial regularizer into the fuzzy objective function. Formulation of the new fuzzy objective function is presented; and its iterative numerical solution, which minimizes the objective function, derived. We applied the proposed algorithm for the segmentation of brain MRI data. Experimental results have demonstrated that the proposed clustering method outperforms other fuzzy clustering models.

Feature weighting plays an important role in improving the performance of clustering technique. We propose an automated feature weighting in fuzzy declustering-based vector quantization (FDVQ), namely AFDVQ algorithm, for enhancing effectiveness and efficiency in classification. The proposed AFDVQ imposes weights on the modified fuzzy c-means (FCM) so that it can automatically calculate feature weights based on their degrees of importance rather than treating them equally. Moreover, the extension of FDVQ and AFDVQ algorithms based on generalized improved fuzzy partitions (GIFP), known as GIFP-FDVQ and GIFP-AFDVQ respectively, are proposed. The experimental results on real data (original and noisy data) and modified data (biased and noisy-biased data) have demonstrated that the proposed algorithms outperformed standard algorithms in classifying clusters especially for biased data.

Computerized tools for automated detection of white matter lesions of the brain in magnetic resonance imaging are very useful for neuroscience researchers to enhance the study of brain-related diseases and their causal associations with other risk factors. We introduce in this paper a fusion approach for identifying white matter lesions in elderly subjects with structural brain tissue changes. The detection methodology is based on image segmentation methods and probabilistic models for membership assignments and fusion. Experimental results on image data of patients show the effectiveness of the proposed approach in comparisons with other detection models.

This paper investigates the affect of variation of patterns in protein profiles to the identification of disease-specific biomarkers. A correlation-based cluster-space transform is applied to mass spectral data for predicting major adverse cardiac events (MACE). Training and testing data are transformed into cluster spaces by correlation distance based clustering, respectively. Data in the testing cluster that falls into a pair of training clusters is classified by a supervised classifier. Experiment results have shown that proteomic spectra of MACE which vary with certain patterns could be separated by the correlation-based clustering. The cluster-space transform allows better classification accuracy than single-clustered class method for separating disease and healthy samples.

Investigation on novel methods for extracting objects of interest in medical images has been an important and challenging area of research in image analysis. In particular, medical images are highly spatially correlated and subject to fuzzy distribution of pixels, we present in this paper a new algorithm for medical image segmentation with special reference to abdominal aortic aneurysm and degraded human brain imaging. Development of the new algorithm is based on the implementation of the theoretic distance matrix with spatial semi-variances.

Scaling behavior inherently exists in fundamental biological structures, and the measure of such an attribute can only be known at a given scale of observation. Thus, the properties of fractals and power-law scaling have become attractive for research in biology and medicine because of their potential for discovering patterns and characteristics of complex biological morphologies. Despite the successful applications of fractals for the life sciences, the quantitative measure of the scale invariance expressed by fractal dimensions is limited in more complex situations, such as for histopathological analysis of tissue changes in disease. In this paper, we introduce the concept of fuzzy scaling and its analysis of a mouse mutant with postnatal brain morphological changes.

Proteomic technology has been found promising for classifying complex diseases that leads to early prediction. However, for effective classification, the extraction of good features that can represent the identities of different classes plays the frontal critical factor for any classification problems. In addition, another major problem associated with pattern recognition is how to effectively handle a large feature space. This paper addresses these two frontal issues for mass spectrometry (MS) classification. We apply the theory of linear predictive coding to extract features and fuzzy vector quantization to reduce the large feature space of MS data. The minimization of the fuzzy scattering matrix in the setting of the fuzzy c-means algorithm provides better grouping for feature classification. The proposed methodology was tested using two MS-based cancer datasets and the results are promising.

Abdominal aortic aneurysm (AAA) is a common disease affecting elderly people and increasing in incidence. The most feared complication of AAA is the rupture of which most will result in death. The AAA involves the excessive dilation of the abdominal aorta in diameter. As a result, open surgery or endoluminal repair is indicated in AAA greater than 55 mm. Currently screening and assessment of AAA can be achieved by either ultrasound or computed tomography (CT) angiography, where the latter imaging technology is the current gold standard. Each AAA is different having varying percentage of thrombus, total volume, luminal volume and calcification all of which are thought to play a critical role for assessing the rupture risk and determining management. Currently measurement of these parameters is based on manual or semi-automatic CT image segmentation - it is time-consuming, inaccurate and becomes unrealistic in clinical practice. The development of an automated method for the segmentation of AAA CT images is therefore demanding. We introduce in this paper a geostatistically constrained fuzzyc-means based algorithm as an automatic and effective segmentation of such images.

Recent advances in biomedicine, pharmacology, and biotechnology open doors to the understanding how diseases are developed at the molecular and physiological level. This gain of understanding tremendously helps facilitate the design and discovery of drugs for therapeutic treatment. Despite the advances in the technology and new knowledge in systems biology, drug discovery is still a low process without utilizing scientific computations that allow precise and rapid analysis of biological processes under trials. This paper particularly addresses fractals as a computational tool for analyzing molecular imaging data that appear to be very useful sources of information for understanding the interactions and behaviors of complex biological networks and the development of predictive medicine. We study herein some fractal characteristics of fluorescent microscope images of peroxisomes, and propose the conceptual frameworks of fuzzy mixture fractal dimensions and fractal distortion measures for bioimage classification.

Early prediction of the risk of cardiovascular events in patients with chest pain is critical in order to provide appropriate medical care for those with positive diagnosis. This paper introduces a computational methodology for predicting such events in the context of robust computerized classification using mass spectrometry data of blood samples collected from patients in emergency departments. We applied the computational theories of statistical and geostatistical linear prediction models to extract effective features of the mass spectra and a simple decision logic to classify disease and control samples for the purpose of early detection. While the statistical and geostatistical techniques provide better results than those obtained from some other methods, the geostatistical approach yields superior results in terms of sensitivity and specificity in various designs of the data set for validation, training, and testing. The proposed computational strategies are very promising for predicting major adverse cardiac events within six months.

Given a newly found gene of some particular genome and a database of sequences whose functions have been known, it must be very helpful if we can search through the database and identify those that are similar to the particular new sequence. The search results may help us to understand the functional role, regulation, and expression of the new gene by the inference from the similar database sequences. This is the task of any methods developed for biological database searching. In this paper we present a new application of the theories of linear predictive coding, vector quantization, and hidden Markov models to address the problem of DNA sequence similarity search where there is no need for sequence alignment. The proposed approach has been tested and compared with some existing methods against real DNA and genomic datasets. The experimental results demonstrate its potential use for such purpose.

Gaussian mixture model (GMM) is used in cell phase identification to model the distribution of cell feature vectors. The model parameters, which are mean vectors, covariance matrices and mixture weights, are trained in an unsupervised learning method using the expectation maximization algorithm. Experiments have shown that the GMM is an effective method capable of achieving high identification rate. However, the GMM approach is not always effective because of ambiguity inherently existing in the cell phase data. To enhance the effectiveness of the GMM for solving this specific problem, the relaxation labeling (RL) is proposed to be used with the GMM. The RL algorithm is a parallel algorithm that updates the probabilities of cell phases by using correlation or mutual information between cell phases to reduce uncertainty among GMMs having overlapping properties.

Advances in fluorescent probing and microscopic imaging technology provide important tools for biomedical research in studying the structures and functions of cells and molecules. Such studies require the processing and analysis of huge amounts of image data, and manual image analysis is very time consuming, thus costly, and also potentially inaccurate and poor reproducibility. In this paper, we present and combine several advanced computational, probabilistic, and fuzzy-set methods for the computerized classification of cell nuclei in different mitotic phases. We tested our proposed methods with real image sequences recorded over a period of twenty-four hours at every fifteen minutes with a time-lapse fluorescence microscopy. The experimental results have shown that the proposed methods are effective for the task of classification.

MOTIVATION: The discrimination and measurement of fluorescent-labeled vesicles using microscopic analysis of fixed cells presents a challenge for biologists interested in quantifying the abundance, size and distribution of such vesicles in normal and abnormal cellular situations. In the specific application reported here, we were interested in quantifying changes to the population of a major organelle, the peroxisome, in cells from normal control patients and from patients with a defect in peroxisome biogenesis. In the latter, peroxisomes are present as larger vesicular structures with a more restricted cytoplasmic distribution. Existing image processing methods for extracting fluorescent cell puncta do not provide useful results and therefore, there is a need to develop some new approaches for dealing with such a task effectively. RESULTS: We present an effective implementation of the fuzzy c-means algorithm for extracting puncta (spots), representing fluorescent-labeled peroxisomes, which are subject to low contrast. We make use of the quadtree partition to enhance the fuzzy c-means based segmentation and to disregard regions which contain no target objects (peroxisomes) in order to minimize considerable time taken by the iterative process of the fuzzy c-means algorithm. We finally isolate touching peroxisomes by an aspect-ratio criterion. The proposed approach has been applied to extract peroxisomes contained in several sets of color images and the results are superior to those obtained from a number of standard techniques for spot extraction. AVAILABILITY: Image data and computer codes written in Matlab are available upon request from the first author.

MOTIVATION: Alignment-free sequence comparison methods are still in the early stages of development compared to those of alignment-based sequence analysis. In this paper, we introduce a probabilistic measure of similarity between two biological sequences without alignment. The method is based on the concept of comparing the similarity/dissimilarity between two constructed Markov models. RESULTS: The method was tested against six DNA sequences, which are the thrA, thrB and thrC genes of the threonine operons from Escherichia coli K-12 and from Shigella flexneri; and one random sequence having the same base composition as thrA from E.coli. These results were compared with those obtained from CLUSTAL W algorithm (alignment-based) and the chaos game representation (alignment-free). The method was further tested against a more complex set of 40 DNA sequences and compared with other existing sequence similarity measures (alignment-free). AVAILABILITY: All datasets and computer codes written in MATLAB are available upon request from the first author.

We apply different advantages of the optimal genetic searching, geostatistics, and fuzzy c-means clustering to the segmentation of gray-level images. The proposed method can deal effectively with noisy image segmentation.

Presents an extended version of the fuzzy c-means algorithm that takes into account the probabilistic information of spatial datasets. This spatial probability can be determined by the indicator kriging of geostatistical estimation. The proposed approach is also considered as the fusion of probabilistic and fuzzy evidences, which are complementary to each other in the data clustering process.

Ordinary kriging and fuzzy sets are combined to derive a spatial filter for restoring degraded images. As kriging is a nonconvex estimation technique and negative kriging weights applied to image data can give estimates outside the range of pixel values. Convexity is therefore required in this image analysis to ensure no negative weights. Fuzzy sets are used to enhance the smoothing process of an ordinary kriging filter. Experiments on an image degraded by Gaussian white noise are given to illustrate the effectiveness of the proposed approach in comparison with the adaptive Wiener filter.

An ordinary kriging based approach for restoring degraded images is presented in this paper. Kriging is a nonconvex estimation technique and negative kriging weights applied to image data can give estimates outside the range of pixel values. Convexity is therefore required in this image analysis to ensure no negative weights. Experiments on an image degraded with different levels of Gaussian white noise are given to illustrate the effectiveness of the proposed approach. As a result, noisy images restored by ordinary kriging filter are more favorable than those restored by the adaptive Wiener filter.

We discuss the conceptual and computational frameworks of information theory for decision making in speaker verification. The proposed approach departs from other conventional scoring models for speaker verification as the first approach takes into account the quantity of 'surprise' or information content. We compare the new approach with a widely used log-likelihood normalization method for speaker verification. Experimental results on a commercial speech corpus validates the theoretical foundation of the proposed method. Furthermore, we introduce the unique entropic measure of uncertainty in the verification scoring.

Formal methods are used to improve the quality of complex computer software by means of documenting system specifications in a precise and structured manner, the most popular specification language for formal methods is Z. However, based on classical set theory and classical logic, this mathematical language can only deal effectively with well-defined problems. This is a disadvantage that classical set operators and classical predicate logic can offer to formal methods. In this paper, the theory of fuzzy information granulation is discussed with an attempt to build toward flexible formal software specifications in which many aspects of human reasoning and natural language can be effectively addressed in mathematical terms. In other words, the tolerance of imprecision necessarily required in many real-life software systems can be represented in the clear and structured mathematics of the fuzzy information granulation theory within the extended framework of formal methods. © 2000 John Wiley & Sons, Inc.

A kriging method is presented as a spatial filter for smoothing gray-scale images degraded by Gaussian white noise. The concepts are based on the analysis of semivariances, the linear combination scheme of kriging, and fuzzy sets. Application of fuzzy sets allows a gradual transition between two boundaries of semivariance levels as a criterion for smoothing the pixel values. This fuzzy thresholding also allows some degree of flexibility to suit various desired results for particular problems. Experimental results obtained by the fuzzy kriging filter are smoother and still preserve edges compared with those by the adaptive Wiener filter.

A face recognition system is described which employs a fuzzy information fusion technique to increase the overall recognition rate. The face images are searched for locating head area and face boundary. The eyes, and mouth are detected using rigid and deformable templates. Assuming a 3D head model, the face rotations are estimated, which allows for compensating rotated facial features back to a front, upright view. Each facial feature forms a source of information for classification. Based on a correlation technique using eye-forehead, mouth and nose windows, three classifiers are established. The output of each classifier is taken as a partial evidence in classification. The importance of each source is measured using a fuzzy density measure and the final classification is achieved using a fuzzy evidence aggregation method. The performance of the system is evaluated using a combined match score.

A nonlinear probabilistic model of the relaxation labeling (RL) process is implemented in the speaker identification task in order to disambiguate the labeling of the speech fea-ture vectors. Identification rates using the RL are higher than those using the conventional VQ (vector quantization) method. 1.

Similarity normalization techniques are important for speaker verification systems as they help to better cope with speaker variability. In the conventional normalization, the a priori probabilities of the cohort speakers are assumed to be equal. From this standpoint, we apply the theory of fuzzy measure and fuzzy integral to combine the likelihood values of the cohort speakers in which the assumption of equal a priori probabilities is relaxed. This approach replaces the conven-tional normalization term by the fuzzy integral which acts as a non-linear fusion of the similarity measures of an ut-terance assigned to cohort speakers. Experimental results show that the speaker verification system using the fuzzy integral is more flexible and favorable than the conventional method. 1.