Visualization and Analysis of Myocardial Transition by Cardiac Ultrasound Images

Abstract

In this paper we present a technique for visualization and analysis of cardiac wall motion, i.e., myocardial transition abnormalities, by digital cardiac ultrasound B-mode images. This technique is based on the correspondence analysis of characteristic points, called landmarks, extracted from endocardial boundary locations and can be used for wall motion analysis of the heart. The algorithm for this technique is that first, cardiac images are segmented into the myocardium and the background image using the gradient relaxation method and endocardial boundary locations are extracted. Next, the landmarks with geometric features are specified, and correspondence analysis is performed with landmarks of the next image. Functional diagnosis of moving organs can be performed by classifying the endocardial boundary into some appropriate colors and superimposing on the original B-mode image.