Segmentation of left ventricle myocardium in MRI using endocardium layer expansion model

We present a volumetric approach for segmentation of left ventricle (LV) myocardium in MRI based on fuzzy connectedness and incorporation of voxel with minimum cost. It is a two-step technique: first, segmentation of LV cavity and then expansion of muscle layers keeping roughly the endocardium morphology. The LV cavity is extracted applying competitive fuzzy connectedness between cavity and myocardium. A small region sample of each object is provided by the user to characterize their membership function that is based on likelihood. From the endocardium surface, the connected voxels with minimum insertion cost and below a calculated threshold are incorporated in an orderly fashion. The cost function encompasses local gradient, local texture and shape.