Jitendra Virmani

Prediction of liver cirrhosis based on multiresolution texture descriptors from B-mode ultrasound

A computer aided diagnostic system to characterise normal and cirrhotic liver by multiresolution texture descriptors is proposed in this paper. The study is carried out in 120 segmented regions of interest extracted from 31 clinically acquired B-mode liver ultrasound images. Mean and standard deviation multiresolution texture descriptors derived by using 2D-discrete wavelet transform, 2D-wavelet packet transform and 2D-Gabor wavelet transform are considered for analysis and exhaustive search with J3 criterion of class separability is used for feature selection. The performance of subset of five most discriminative texture descriptors obtained from 2D-discrete wavelet transform, 2D-wavelet packet transform and 2D-Gabor wavelet transform is compared by using a support vector machine classifier. It is observed that only five mean multiresolution texture descriptors obtained from 2D-Gabor wavelet transform at selective scale and orientations provide highest classification accuracy of 98.33% and sensitivity of 100% by using a support vector machine classifier. The promising results indicate that the selective frequency and orientation properties of Gabor filters are extremely useful for providing multiscale texture description.

A Decision Support System for Classification of Normal and Medical Renal Disease Using Ultrasound Images: A Decision Support System for Medical Renal Diseases

Early detection of medical renal disease is important as the same may lead to chronic kidney disease which is an irreversible stage. The present work proposes an efficient decision support system for detection of medical renal disease using small feature space consisting of only second order GLCM statistical features computed from raw renal ultrasound images. The GLCM mean feature vector and GLCM range feature vector are computed for inter-pixel distance d varying from 1 to 10. These texture feature vectors are combined in various ways yielding GLCM ratio feature vector, GLCM additive feature vector and GLCM concatenated feature vector. The present work explores the potential of five texture feature vectors computed using GLCM statistics exhaustively for differential diagnosis between normal and MRD images using SVM classifier. The result of the study indicates that GLCM range feature vector computed with d = 1 yields the highest overall classification accuracy of 85.7% with individual classification accuracy values of 93.3% and 77.9% for normal and MRD classes respectively.

Breast Tissue Density Classification Using Wavelet-Based Texture Descriptors

It has been well established that the risk of breast cancer development is associated with increased breast density. Therefore, characterization of breast tissue density is clinically significant. In the present work, the potential of various wavelet energy descriptors (derived from ten different compact support wavelet filters) has been investigated for breast tissue density classification using kNN, SVM, and PNN classifiers. The work has been carried out on the MIAS dataset. The highest classification accuracy of 96.2 % is achieved using the kNN classifier Haar wavelet energy descriptors.

Detection of Chronic Kidney Disease: A NN-GA-Based Approach

In the present work, a genetic algorithm (GA) trained neural network (NN)-based model has been proposed to detect chronic kidney disease (CKD) which has become one of the newest threats to the developing and undeveloped countries. Studies and surveys in different parts of India have suggested that CKD is becoming a major concern day by day. The financial burden of the treatment and future consequences of CKD could be unaffordable to many, if not detected at an earlier stage. Motivated by this, the NN-GA model has been proposed which significantly overcomes the problem of using local search-based learning algorithms to train NNs. The input weight vector of the NN is gradually optimized by using GA to train the NN. The model has been compared with well-known classifiers like Random Forest, Multilayer Perception Feedforward Network (MLP-FFN), and also with NN. The performance of the classifiers has been measured in terms of accuracy, precision, recall, and F-Measure. The experimental results suggest that NN-GA-based model is capable of detecting CKD more efficiently than any other existing model.

Comparison of CAD Systems for Three Class Breast Tissue Density Classification Using Mammographic Images

It is well known that the changes in the breast tissue density are strongly correlated with the risk of breast cancer development and therefore classifying the breast tissue density as fatty, fatty–glandular and dense–glandular has become clinically significant. It is believed that the changes in the tissue density can be captured by computing the texture descriptors. Accordingly, the present work has been carried out with an aim to explore the potential of Laws’ mask texture descriptors for description of variations in breast tissue density using mammographic images. The work has been carried out on the 322 mammograms taken from the MIAS dataset. The dataset consists of 106 fatty, 104 fatty–glandular and 112 dense–glandular images. The ROIs of size 200 × 200 pixels are extracted from the center of the breast tissue, ignoring the pectoral muscle. For the design of a computer aided diagnostic system for three class breast tissue density classification, Laws’ texture descriptors have been computed using Laws’ masks of different resolutions. Five statistical features i.e. mean, skewness, standard deviation, entropy and kurtosis have been computed from all the Laws’ texture energy images generated from each ROI. The feature space dimensionality reduction has been carried out by using principal component analysis. For the classification task kNN, PNN and SVM classifiers have been used. After carrying out exhaustive experimentation, it has been observed that PCA–SVM based CAD system design yields the highest overall classification accuracy of 87.5 %, with individual class accuracy values of 84.9, 84.6 and 92.8 % for fatty, fatty–glandular and dense–glandular image classes respectively. These results indicate the usefulness of the proposed CAD system for breast tissue density classification.

Application of Texture Features for Classification of Primary Benign and Primary Malignant Focal Liver Lesions

The present work focuses on the aspect of textural variations exhibited by primary benign and primary malignant focal liver lesions. For capturing these textural variations of benign and malignant liver lesions, texture features are computed using statistical methods, signal processing based methods and transform domain methods. As an application of texture description in medical domain, an efficient CAD system for primary benign i.e., hemangioma (HEM) and primary malignant i.e., hepatocellular carcinoma (HCC) liver lesions based on texture features derived from B-Mode liver ultrasound images of Focal liver lesions has been proposed in the present study. The texture features have been computed from the inside regions of interest (IROIs) i.e., from the regions inside the lesion and one surrounding region of interest (SROI) for each lesion. Texture descriptors are computed from IROIs and SROIs using six feature extraction methods namely, FOS, GLCM, GLRLM, FPS, Gabor and Laws’ features. Three texture feature vectors (TFVs) i.e., TFV1 consists of texture features computed from IROIs, TFV2 consists of texture ratio features (i.e., texture feature value computed from IROI divided by texture feature value computed from corresponding SROI) and TFV3 computed by combining TFV1 and TFV2 (IROIs texture features + texture ratio features) are subjected to classification by SVM and SSVM classifiers. It is observed that the performance of SSVM based CAD system is better than SVM based CAD system with respect to (a) overall classification accuracy (b) individual class accuracy for atypical HEM class and (c) computational efficiency. The promising results obtained from the proposed SSVM based CAD system design indicates its usefulness to assist radiologists for differential diagnosis between primary benign and primary malignant liver lesions.

Application of Statistical Texture Features for Breast Tissue Density Classification

It has been strongly advocated that increase in density of breast tissue is strongly correlated with the risk of developing breast cancer. Accordingly change in breast tissue density pattern is taken seriously by radiologists. In typical cases, the breast tissue density patterns can be easily classified into fatty, fatty-glandular and dense glandular classes, but the differential diagnosis between atypical breast tissue density patterns from mammographic images is a daunting challenge even for the experienced radiologists due to overlap of the appearances of the density patterns. Therefore a CAD system for the classification of the different breast tissue density patterns from mammographic images is highly desirable. Accordingly in the present work, exhaustive experiments have been carried out to evaluate the performance of statistical features using PCA-kNN, PCA-PNN, PCA-SVM and PCA-SSVM based CAD system designs for two-class and three-class breast tissue density classification using mammographic images. It is observed that for two-class breast tissue density classification, the highest classification accuracy of 94.4 % is achieved using only the first 10 principal components (PCs) derived from statistical features with the SSVM classifier. For three-class breast tissue density classification, the highest classification accuracy of 86.3 % is achieved using only the first 4 PCs with SVM classifier.

Classification of Breast Tissue Density Patterns Using SVM-Based Hierarchical Classifier

In the present work, three-class breast tissue density classification has been carried out using SVM-based hierarchical classifier. The performance of Laws’ texture descriptors of various resolutions have been investigated for differentiating between fatty and dense tissues as well as for differentiation between fatty-glandular and dense-glandular tissues. The overall classification accuracy of 88.2% has been achieved using the proposed SVM-based hierarchical classifier.