Fast Level Set-based Segmentation without Edge and Re-Initialization

H. Zhou, C. Yuan, and Y. Liu (PRC)

Keywords

Active contours, without edges, re-initialization, level sets, segmentation, PDE

Abstract

Nowadays Level set method base on partial differential equations(PDE) become a popular tool for image segmen tation. Active contours without edges proposed by Chan Vese is prone to global minimal but time-consuming. Level set evolution without re-initialization method is base on gradient of the image, so it is likely to converge to the local minimal. In this paper, we introduced Active Contours without Edges and Level Set Evolution without Reinitialization, then present a variational formulation which is faster than both of the method mentioned above. The proposed method has two main advantages over the traditional level set methods. First, this method has a significantly speeds up the curve evolution. Second, the initialize of the level set function can be arbitrarily selected. Moreover, our method can be easily implemented by simple finite difference scheme. We apply our method on some images and compare with traditional method in a table.

Important Links:



Go Back