Active contours under topology control genus preserving level sets

摘要: We present a novel framework to exert topology control over level set evolution. Level methods offer several advantages parametric active contours, in particular automated topological changes. In some applications, where priori knowledge of the target is available, changes may not be desirable. This typically case biomedical image segmentation, shape prescribed by anatomical knowledge. However, topologically constrained evolutions often generate barriers that lead large geometric inconsistencies. introduce controlled greatly alleviates this problem. Unlike existing work, our method allows connected components merge, split or vanish under specific conditions ensure no defects are generated. demonstrate strength on wide range numerical experiments and illustrate its performance segmentation cortical surfaces blood vessels.