Measurements of texture features of medical images and its application to computer-aided diagnosis in cardiomyopathy

In this paper we use four texture features of ultrasonic images for classification of the heart disease. The texture features are measured from a gray-level cooccurrence matrix generated from the heart images and are used as inputs to our classification system incorporated with soft computing technologies, which include fuzzy logic-based systems, neural-network-based systems and genetic algorithms. We evaluate the performance of our proposed method in terms of sensitivity, specificity and overall accuracy. Experimental results show that the proposed method exploiting the texture features has potential utility for computer-aided diagnosis of myocardial heart disease.