标题：Two new fast algorithms for variational multiphase image segmentation
作者：Liu, Cunliang ;Pan, Zhenkuan ;Zheng, Yongguo ;Wang, Guodong ;Duan, Jinming
作者机构：[Liu, Cunliang ;Zheng, Yongguo ] College of Information Science and Engineering, Shandong University of Science and Technology, Qingdao 266510, China; 更多
来源：International Journal of Digital Content Technology and its Applications
摘要：The problems of image segmentation are often solved by the level set method, which generally leads to solutions of some gradient descent equations with low computation efficiency. In this paper, we present an augmented Lagrangian method and a direct dual method to accelerate the implementation. For the augmented Lagrangian method, we incorporate the variable splitting approach and Uzawa scheme to update the Lagrange multiplier. For the direct dual method, we use the KKT condition to get a binary analytical expression for the primal variable without convex relaxation. We apply the two approaches to the multiphase Chan-Vese model and obtain two efficient alternative iterative algorithms. Numerical results on synthetic and real images are provided to compare our methods with some others, which demonstrate effectiveness and efficiency of our algorithms.