Snake Model-Based Automatic Segmentation of the Left Ventricle from Cardiac MR Images

An approach based on selective smoothing direction gradient vector flow (SSDGVF) snake model incorporating shape prior is proposed to segment the left ventricle from cardiac MR images in this paper. The originalities of the presented method include SSDGVF algorithm, automatic localization of the cardiac endocardium contour, and elliptic shape constraint. This novel approach can overcome the unexpected local minimum, and conquer the weak boundary leakage in tracking the boundaries of the left ventricle myocardium. Validation is performed on a set of 21 cardiac MR images, and satisfactory segmentation results are obtained.