Segmentation and tracking of the cardiac muscle in echocardiographic sequences using a level set approach

In the context of echocardiographic imaging, we propose a level set formulation for registration and segmentation with a shape prior. Registration of a prior contour on the image is expressed as the problem of estimating a rigid transform minimizing a global region-based criterion. The segmentation step is then formulated through front propagation constrained with the registered shape prior. Based on the result of segmentation, the motion field of the segmented muscle is derived and is then used to track a region through the cardiac cycle. This segmentation and tracking framework was applied to in vivo data and was shown to provide very consistent results.