Mitral regurgitation assessment by automated analysis of intracavitary contrast 2-D echocardiograms

摘要

An automated system based on a PC/AT-286 clone with image acquisition and processing boards is proposed for evaluating mitral regurgitation (MR) through a series of analysis of the intracavitary contrast two-dimensional echocardiograms. Four MR classes, classified according to the regurgitation severity degree, have been studied. Features from these four classes have been extracted in a training set of 79 randomly selected samples. Potentially significant features were selected for a hierarchical linear classifier. A total rate of 75% correct classification was obtained which illustrates the efficiency of the approach and its potential in clinical practice.