A New Method for Endocardial Boundary Detection Using Genetic Algorithms

摘要

This paper presents a new method for endocardial boundary detection that consists of two major parts. First, the initial endocardial boundary extraction process is designed to reduce the impact of signal dropout in 2-D echocardiograms. Second, the detected endocardial boundary is then refined by the use of genetic algorithms. The proposed refinement step takes the advantages of population-wide searching and energy convergence capability in genetic algorithms. The near-optimal (or optimal) and stable results can be obtained for the detection of cardiac boundary. Experiments on several left ventricular (LV) echocardiograms have been clinically verified for its effectiveness and convenience in handling patient data.