A novel proximity algorithm for an unconstrained convex segmentation model
Weibin Li, Yanxia Du, Xian Yi, Honglin Ma, Chao Song · 2019
A new proximity algorithm for an active contour model is proposed in this paper. In order to derive a mathematical form of the energy, the level set method is employed. In the new energy, a penalty term is introduced to make sure that the level set function can be restricted in the interval [-1, 1]. By introducing this term, the energy still keeps convex and is easy to construct its minimization algorithms. Based on the proximity operator and the corresponding theories, we deduce a proximity algorithm to minimize the energy. Experimental results demonstrate that the proposed model is powerful in its segmentation ability and accuracy. And comparisons with other popular algorithms show that the proposed algorithm is more efficient.