Computerized analysis for classification of heart diseases in echocardiographic images
Du‐Yih Tsai, Satoru Watanabe, Masaaki Tomita · 2002
The classification of ultrasonic heart (echocardiographic) images is studied by making use of some texture features, including the angular second moment, contrast, correlation and entropy which are obtained from a gray-level cooccurrence matrix. Features of these types are used as inputs to the input layer of a neural network (NN) to classify three sets of echocardiographic images: normal heart, dilated cardiomyopathy (DCM) and hypertrophic cardiomyopathy (HCM) (18, 13 and 10 samples, respectively). The performance of the NN classifier is compared to that of decision-theoretic (D-T) method. Moreover, fractal dimension (FD) is also used as an independent feature of images for discrimination. Our results show that the NN, D-T method and FD method produce about 94%, 89%, and 84% correct classification, respectively. The results indicate that the method of feature-based image analysis using the NN has potential utility for computer-aided diagnosis of the DCM, HCM and other heart diseases.