U. Raghavendra

Automated characterization of fatty liver disease and cirrhosis using curvelet transform and entropy features extracted from ultrasound images

Fatty liver disease (FLD) is reversible disease and can be treated, if it is identified at an early stage. However, if diagnosed at the later stage, it can progress to an advanced liver disease such as cirrhosis which may ultimately lead to death. Therefore, it is essential to detect it at an early stage before the disease progresses to an irreversible stage. Several non-invasive computer-aided techniques are proposed to assist in the early detection of FLD and cirrhosis using ultrasound images. In this work, we are proposing an algorithm to discriminate automatically the normal, FLD and cirrhosis ultrasound images using curvelet transform (CT) method. Higher order spectra (HOS) bispectrum, HOS phase, fuzzy, Kapoor, max, Renyi, Shannon, Vajda and Yager entropies are extracted from CT coefficients. These extracted features are subjected to locality sensitive discriminant analysis (LSDA) feature reduction method. Then these LSDA coefficients ranked based on F-value are fed to different classifiers to choose the best performing classifier using minimum number of features. Our proposed technique can characterize normal, FLD and cirrhosis using probabilistic neural network (PNN) classifier with an accuracy of 97.33%, specificity of 100.00% and sensitivity of 96.00% using only six features. In addition, these chosen features are used to develop a liver disease index (LDI) to differentiate the normal, FLD and cirrhosis classes using a single number. This can significantly help the radiologists to discriminate FLD and cirrhosis in their routine liver screening.

Application of Gabor wavelet and Locality Sensitive Discriminant Analysis for automated identification of breast cancer using digitized mammogram images

Display Omitted Classification of normal, benign and malignant mammograms is proposed.Gabor wavelet coupled with Locality Sensitive Discriminant Analysis is used.Achieved accuracy of 98.69% for kNN classifier with eight features. Breast cancer is one of the prime causes of death in women. Early detection may help to improve the survival rate to a great extent. Mammography is considered as one of the most reliable methods to prescreen of breast cancer. However, reading the mammograms by radiologists is laborious, taxing, and prone to intra/inter observer variability errors. Computer Aided Diagnosis (CAD) helps to obtain fast, consistent and reliable diagnosis. This paper presents an automated classification of normal, benign and malignant breast cancer using digitized mammogram images. The proposed method used Gabor wavelet for feature extraction and Locality Sensitive Discriminant Analysis (LSDA) for data reduction. The reduced features are ranked using their F-values and fed to Decision Tree (DT), Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA), k-Nearest Neighbor (k-NN), Naive Bayes Classifier (NBC), Probabilistic Neural Network (PNN), Support Vector Machine (SVM), AdaBoost and Fuzzy Sugeno (FSC) classifiers one by one to select the highest performing classifier using minimum number of features. The proposed method is evaluated using 690 mammogram images taken from a benchmarked Digital Database for Screening Mammography (DDSM) dataset. Our developed method has achieved mean accuracy, sensitivity, specificity of 98.69%, 99.34% and 98.26% respectively for k-NN classifier using eight features with 10-fold cross validation. This system can be employed in hospitals and polyclinics to aid the clinicians to cross verify their manual diagnosis.

Decision support system for fatty liver disease using GIST descriptors extracted from ultrasound images

Steatosis or fatty liver disease (FLD) is characterized by the abnormal retention of large vacuoles of neutral fat in the liver cells, either due to alcoholism or metabolic syndrome. Succession of FLD can lead to severe liver diseases such as hepatocellular carcinoma, cirrhosis and hepatic inflammation but it is a reversible disease if diagnosed early. Thus, computer-aided diagnostic tools play a very important role in the automated diagnosis of FLD. This paper focuses on the detection of steatosis and classification of steatotic livers from the normal using ultrasound images. The significant information from the image is extracted using GIST descriptor models. Marginal Fisher Analysis (MFA) integrated with Wilcoxon signed-rank test helps to eliminate the trivial features and provides the distinctive features for qualitative classification. Finally the clinically significant features are fused using classifiers such as decision tree (DT), support vector machine (SVM), adaBoost, k-nearest neighbor (kNN), probabilistic neural network (PNN), naive Bayes (NB), fuzzy Sugeno (FS), linear and quadratic discriminant analysis classification of normal and abnormal liver images. Results portray that PNN classifier can diagnose FLD with an average classification accuracy of 98%, 96% sensitivity, 100% specificity and Area Under Curve (AUC) of 0.9674 correctly.

A review on automatic detection and recognition of traffic sign

Evidently, Intelligent Transport System (ITS) has progressed tremendously all its way. The core of ITS are detection and recognition of traffic sign, which are designated to fulfill safety and comfort needs of driver. This paper provides a critical review on three major steps in Automatic Traffic Sign Detection and Recognition(ATSDR) system i.e., segmentation, detection and recognition in the context of vision based driver assistance system. In addition, it focuses on different experimental setups of image acquisition system. Further, discussion on possible future research challenges is made to make ATSDR more efficient, which inturn produce a wide range of opportunities for the researchers to carry out the detailed analysis of ATSDR and to incorporate the future aspects in their research.

Multiple thresholding and subspace based approach for detection and recognition of traffic sign

Automatic detection and recognition of traffic sign has been a topic of great interest in advanced driver assistance system. It enhances vehicle and driver safety by providing the condition and state of the road to the drivers. However, visual occlusion and ambiguities in the real-world scenario make the traffic sign recognition a challenging task. This paper presents an Automatic Traffic Sign Detection and Recognition (ATSDR) system, involving three modules: segmentation, detection, and recognition. Region of Interest (ROI) is extracted using multiple thresholding schemes with a novel environmental selection strategy. Then, the traffic sign detection is carried out using correlation computation between log-polar mapped inner regions and the reference template. Finally, recognition is performed using Support Vector Machine (SVM) classifier. Our proposed system achieved a recognition accuracy of 98.3 % and the experimental results demonstrates the robustness of traffic sign detection and recognition in real-world scenario.

Local texture patterns for traffic sign recognition using higher order spectra

A novel traffic sign recognition module is presented.Higher order spectra and texture based features are extracted.Achieved recognition accuracy of 98.89% using k-NN classifier. Traffic sign recognition (TSR) is considered as one of the most important modules of driver assistance system (DAS). It can be used as a decision supporting tool for driver and autonomous vehicles. Eventually, TSR is a large-scale feature learning problem and hence attracted the attention of researchers recently. The essential parameters such as huge training dataset size, recognition accuracy and computational complexity need to be considered while designing a practical TSR system. In this paper, we have used higher order spectra (HOS) coupled with texture based features to develop an efficient TSR model. These features represent the shape and content of the traffic signs clearly. Then a subspace learning method with graph embedding under linear discriminant analysis framework is used to increase the discrimination power between various traffic symbols. As a result the proposed method attained a maximum recognition accuracy of 98.89%. The proposed method is evaluated using two publicly available datasets such as, Belgium traffic sign classification (BTSC) and German traffic sign recognition benchmark (GTSRB). Our experimental results demonstrate that the proposed approach is computationally efficient and shows promising recognition accuracy.

Diagnosis of retinal health in digital fundus images using continuous wavelet transform (CWT) and entropies

Vision is paramount to humans to lead an active personal and professional life. The prevalence of ocular diseases is rising, and diseases such as glaucoma, Diabetic Retinopathy (DR) and Age-related Macular Degeneration (AMD) are the leading causes of blindness in developed countries. Identifying these diseases in mass screening programmes is time-consuming, labor-intensive and the diagnosis can be subjective. The use of an automated computer aided diagnosis system will reduce the time taken for analysis and will also reduce the inter-observer subjective variabilities in image interpretation. In this work, we propose one such system for the automatic classification of normal from abnormal (DR, AMD, glaucoma) images. We had a total of 404 normal and 1082 abnormal fundus images in our database. As the first step, 2D-Continuous Wavelet Transform (CWT) decomposition on the fundus images of two classes was performed. Subsequently, energy features and various entropies namely Yager, Renyi, Kapoor, Shannon, and Fuzzy were extracted from the decomposed images. Then, adaptive synthetic sampling approach was applied to balance the normal and abnormal datasets. Next, the extracted features were ranked according to the significances using Particle Swarm Optimization (PSO). Thereupon, the ranked and selected features were used to train the random forest classifier using stratified 10-fold cross validation. Overall, the proposed system presented a performance rate of 92.48%, and a sensitivity and specificity of 89.37% and 95.58% respectively using 15 features. This novel system shows promise in detecting abnormal fundus images, and hence, could be a valuable adjunct eye health screening tool that could be employed in polyclinics, and thereby reduce the workload of specialists at hospitals.

Fusion of spatial gray level dependency and fractal texture features for the characterization of thyroid lesions

&NA; Thyroid is a small gland situated at the anterior side of the neck and one of the largest glands of the endocrine system. The abrupt cell growth or malignancy in the thyroid gland may cause thyroid cancer. Ultrasound images distinctly represent benign and malignant lesions, but accuracy may be poor due to subjective interpretation. Computer Aided Diagnosis (CAD) can minimize the errors created due to subjective interpretation and assists to make fast accurate diagnosis. In this work, fusion of Spatial Gray Level Dependence Features (SGLDF) and fractal textures are used to decipher the intrinsic structure of benign and malignant thyroid lesions. These features are subjected to graph based Marginal Fisher Analysis (MFA) to reduce the number of features. The reduced features are subjected to various ranking methods and classifiers. We have achieved an average accuracy, sensitivity and specificity of 97.52%, 90.32% and 98.57% respectively using Support Vector Machine (SVM) classifier. The achieved maximum Area Under Curve (AUC) is 0.9445. Finally, Thyroid Clinical Risk Index (TCRI) a single number is developed using two MFA features to discriminate the two classes. This prototype system is ready to be tested with huge diverse database. HighlightsAutomated system for classification of benign and malignant thyroid lesions.SGLDF and fractal textures are coupled with MFA is used.242 ultrasound images are used for the study.Classification accuracy of 97.52% is obtained using SVM classifier.Formulated TCRI to discriminate the two classes using one integer.

Automated screening of congestive heart failure using variational mode decomposition and texture features extracted from ultrasound images

Heart is an important and hardest working muscular organ of the human body. Inability of the heart to restore normal perfusion to the entire body refers to cardiac failure, which then with symptoms results in manifestation of congestive heart failure (CHF). Impairment in systolic function associated with chronic dilation of left ventricle is referred as dilated cardiomyopathy (DCM). The clinical examination, surface electrocardiogram (ECG), chest X-ray, blood markers and echocardiography play major role in the diagnosis of CHF. Though the ECG manifests chamber enlargement changes, it does not possess sensitive marker for the diagnosis of DCM, whereas echocardiographic assessment can effectively reveal the presence of asymptomatic DCM. This work proposes an automated screening method for classifying normal and CHF echocardiographic images affected due to DCM using variational mode decomposition technique. The texture features are extracted from variational mode decomposed image. These features are selected using particle swarm optimization and classified using support vector machine classifier with different kernel functions. We have validated our experiment using 300 four-chamber echocardiography images (150: normal, 150: CHF) obtained from 50 normal and 50 CHF patients. Our proposed approach yielded maximum average accuracy, sensitivity and specificity of 99.33%, 98.66% and 100%, respectively, using ten features. Thus, the developed diagnosis system can effectively detect CHF in its early stage using ultrasound images and aid the clinicians in their diagnosis.

Automated technique for coronary artery disease characterization and classification using DD-DTDWT in ultrasound images

Abstract Heart is one of the important as well as hardest working organ of human body. Cardiac ischemia is the condition where sufficient blood and oxygen will not reach the heart muscle due to narrowed arteries of the heart. This condition is called coronary artery disease. Several non-invasive diagnostic tests fail to reveal exact impact of coronary artery disease on myocardial segments. The ultrasound images can explore major impact on ventricular muscle segments due to ischemia and complication of acute coronary syndrome. Computer aided diagnosis tools can predict coronary artery disease in its early stage so that patients can undergo treatment and save their life. This paper presents a novel computer aided diagnosis system for the automated detection of coronary artery disease using echocardiography images taken from four chamber heart. Proposed method uses double density-dual tree discrete wavelet transform (DD-DTDWT) to decompose the images into different frequency sub-bands. Then various entropy features are extracted from these sub-bands. The obtained dimension of the features is reduced using marginal fisher analysis (MFA) and optimal features are selected using feature ranking methods. The proposed method achieved promising accuracy of 96.05%, sensitivity of 96.12%, and specificity of 96.00% for linear discriminant classifier using entropy ranking method with twelve features. We have also proposed coronary artery disease risk index (CADRI) to categorize diseased subjects from normal subjects using a single value. Thus, it can be used as a diagnosis tool in hospitals and polyclinics for confirming the findings of clinicians.