An improved fast watershed algorithm for segmentation of echocardiographic images

Echocardiography with delineated ventricle borders is a principal technique for quantitative assessment of cardiac function. As it is, most of the prevailing segmentation methods suffer in detecting edges at equivocal regions. This paper presents an effective method for tracking Left Ventricle borders with a lower time complexity. We propose an improved fast watershed algorithm (IFWSA), which scans the image only twice, one for generating connected components and other for labeling the catchment basins. Experiments over 200 echocardiographic images show that the proposed method ascertains 98% of acceptance level for tracked left ventricle when compared with expert drawn curve, twice faster than existing algorithm.