Sumeet Dua

Classification of heart rate data using artificial neural network and fuzzy equivalence relation

The electrocardiogram is a representative signal containing information about the condition of the heart. The shape and size of the P-QRS-T wave, the time intervals between its various peaks, etc. may contain useful information about the nature of disease afflicting the heart. However, these subtle details cannot be directly monitored by the human observer. Besides, since bio-signals are highly subjective, the symptoms may appear at random in the time scale. Therefore, the signal parameters, extracted and analysed using computers, are highly useful in diagnostics. This paper deals with the classification of certain diseases using artificial neural network (ANN) and fuzzy equivalence relations. The heart rate variability is used as the base signal from which certain parameters are extracted and presented to the ANN for classification. The same data is also used for fuzzy equivalence classifier. The feedforward architecture ANN classifier is seen to be correct in about 85% of the test cases, and the fuzzy classifier yields correct classification in over 90% of the cases.

Automated Diagnosis of Glaucoma Using Texture and Higher Order Spectra Features

Glaucoma is the second leading cause of blindness worldwide. It is a disease in which fluid pressure in the eye increases continuously, damaging the optic nerve and causing vision loss. Computational decision support systems for the early detection of glaucoma can help prevent this complication. The retinal optic nerve fiber layer can be assessed using optical coherence tomography, scanning laser polarimetry, and Heidelberg retina tomography scanning methods. In this paper, we present a novel method for glaucoma detection using a combination of texture and higher order spectra (HOS) features from digital fundus images. Support vector machine, sequential minimal optimization, naive Bayesian, and random-forest classifiers are used to perform supervised classification. Our results demonstrate that the texture and HOS features after z-score normalization and feature selection, and when combined with a random-forest classifier, performs better than the other classifiers and correctly identifies the glaucoma images with an accuracy of more than 91%. The impact of feature ranking and normalization is also studied to improve results. Our proposed novel features are clinically significant and can be used to detect glaucoma accurately.

Wavelet-Based Energy Features for Glaucomatous Image Classification

Texture features within images are actively pursued for accurate and efficient glaucoma classification. Energy distribution over wavelet subbands is applied to find these important texture features. In this paper, we investigate the discriminatory potential of wavelet features obtained from the daubechies (db3), symlets (sym3), and biorthogonal (bio3.3, bio3.5, and bio3.7) wavelet filters. We propose a novel technique to extract energy signatures obtained using 2-D discrete wavelet transform, and subject these signatures to different feature ranking and feature selection strategies. We have gauged the effectiveness of the resultant ranked and selected subsets of features using a support vector machine, sequential minimal optimization, random forest, and naïve Bayes classification strategies. We observed an accuracy of around 93% using tenfold cross validations to demonstrate the effectiveness of these methods.

An integrated diabetic index using heart rate variability signal features for diagnosis of diabetes

Electrocardiogram (ECG) signals are difficult to interpret, and clinicians must undertake a long training process to learn to diagnose diabetes from subtle abnormalities in these signals. To facilitate these diagnoses, we have developed a technique based on the heart rate variability signal obtained from ECG signals. This technique uses digital signal processing methods and, therefore, automates the detection of diabetes from ECG signals. In this paper, we describe the signal processing techniques that extract features from heart rate (HR) signals and present an analysis procedure that uses these features to diagnose diabetes. Through statistical analysis, we have identified the correlation dimension, Poincaré geometry properties (SD2), and recurrence plot properties (REC, DET, L mean) as useful features. These features differentiate the HR data of diabetic patients from those of patients who do not have the illness, and have been validated by using the AdaBoost classifier with the perceptron weak learner (yielding a classification accuracy of 86%). We then developed a novel diabetic integrated index (DII) that is a combination of these nonlinear features. The DII indicates whether a particular HR signal was taken from a person with diabetes. This index aids the automatic detection of diabetes, thereby allowing a more objective assessment and freeing medical professionals for other tasks.

Design and implementation of a unique blood-vessel detection algorithm towards early diagnosis of diabetic retinopathy

Diabetic retinopathy (DR), a major complication of diabetes and the leading cause of new cases of blindness among adults, can be cured by the early and precise detection of the disease. An important aspect of DR is the micro-vascular changes that cause detectable changes in the appearance of retinal blood vessels. In this paper, we propose a new blood-vessel detection technique in retinal images, based on the regional recursive hierarchical decomposition using quadtrees and post-filtration of edges. We exploit the fact that in retinal images, the blood vessels appear as focal and/or penumbral blurred edges, which can be characterized by an estimable intensity gradient, which also serves in dismissing false alarms to a large extent. Our technique provides information on retinal blood vessel morphology that can be calibrated to normal expected blood vessel diameters and which can detect fine blood vessel anomalies that characterize the blood vessel pathology and hence aid early detection of diabetic retinopathy.

Classification of Epilepsy Using High-Order Spectra Features and Principle Component Analysis

The classification of epileptic electroencephalogram (EEG) signals is challenging because of high nonlinearity, high dimensionality, and hidden states in EEG recordings. The detection of the preictal state is difficult due to its similarity to the ictal state. We present a framework for using principal components analysis (PCA) and a classification method for improving the detection rate of epileptic classes. To unearth the nonlinearity and high dimensionality in epileptic signals, we extract principal component features using PCA on the 15 high-order spectra (HOS) features extracted from the EEG data. We evaluate eight classifiers in the framework using true positive (TP) rate and area under curve (AUC) of receiver operating characteristics (ROC). We show that a simple logistic regression model achieves the highest TP rate for class “preictal” at 97.5% and the TP rate on average at 96.8% with PCA variance percentages selected at 100%, which also achieves the most AUC at 99.5%.

Segmentation of fluorescence microscopy cell images using unsupervised mining.

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

Discovery of Web frequent patterns and user characteristics from Web access logs: a framework for dynamic Web personalization

An automatic discovery method that discovers frequent access routines for unique clients from Web access log files is presented. The proposed algorithm develops novel techniques to extract the sets of all predictive access sequences from semi-structured Web access logs. Important user access patterns are manifested through the frequent traversal paths, thus helping to understand user surfing behaviors. The predictive access routines discovered by AllFreSeq are also useful for understanding and improving Web site domain tree.

Thyroid lesion classification in 242 patient population using Gabor transform features from high resolution ultrasound images

Total of 242 benign and malignant thyroid nodules are classified.Various entropies are extracted from Gabor transformed images.These features are subjected to LSDA and ranked by Relief-F method.Various sampling strategies are used to balance the classification data.Obtained classification accuracy of 94.3% with C4.5 decision tree classifier. Thyroid cancer commences from an atypical growth of thyroid tissue at the edge of the thyroid gland. Initially, it forms a lump in the throat and an over-growth of this tissue leads to the formation of benign or malignant thyroid nodules. Blood test and biopsies are the standard techniques used to diagnose the presence of thyroid nodules. But imaging modalities can improve the diagnosis and are marked as cost-effective, non-invasive and risk-free to identify the stages of thyroid cancer. This study proposes a novel automated system for classification of benign and malignant thyroid nodules. Raw images of thyroid nodules recorded using high resolution ultrasound (HRUS) are subjected to Gabor transform. Various entropy features are extracted from these transformed images and these features are reduced by locality sensitive discriminant analysis (LSDA) and ranked by Relief-F method. Over-sampling strategies with Wilcoxon signed-rank, Friedmans and Iman-Davenport post hoc tests are used to balance the classification data and also to improve the classification performance. Classifiers such as support vector machine (SVM), k-nearest neighbour (kNN), multi-layered perceptron (MLP) and decision tree are used for the characterization of benign and malignant thyroid nodules. We have obtained a classification accuracy of 94.3% with C4.5 decision tree classifier using 242 thyroid HRUS images. Our developed system can be used to screen the thyroid automatically and assist the radiologists.

NOVEL CLASSIFICATION OF CORONARY ARTERY DISEASE USING HEART RATE VARIABILITY ANALYSIS

Coronary artery disease (CAD) is a leading cause of death worldwide. Heart rate variability (HRV) has been proven to be a non-invasive marker of the autonomic modulation of the heart. Nonlinear analyses of HRV signals have shown that the HRV is reduced significantly in patients with CAD. Therefore, in this work, we extracted nonlinear features from the HRV signals using the following techniques: recurrence plots (RP), Poincare plots, and detrended fluctuation analysis (DFA). We also extracted three types of entropy, namely, Shannon entropy (ShanEn), approximation entropy (ApEn), and sample entropy (SampEn). These features were subjected to principal component analysis (PCA). The significant principal components were evaluated using eight classification techniques, and the performances of these techniques were evaluated to determine which presented the highest accuracy in classifying normal and CAD classes. We observed that the multilayer perceptron (MLP) method resulted in the highest classification accuracy (89.5%) using our proposed technique.