摘要： 在Li模型的基础上引入C-V模型外部能量项重新构造能量函数，给出一种结合区域与边缘信息的变分水平集模型，结合基于图割理论的GCBAC算法，提出一种图割与改进变分水平集结合的目标提取方法。该方法能够让2种模型有机结合达到优势互补的效果。实验结果表明，该方法具有快速、鲁棒、抗噪性强等优点。

关键词: 目标提取, 变分水平集, 图割, GCBAC算法

Abstract: This paper reconstructs energy function by introducing external energy term of C-V model based on Li model. A variation level set model combining region with edge information is described, and it combines with the graph cuts theory based Graph Cuts Based Active Contours(GCBAC) algorithm. It proposes a method of object extraction which combines graph cuts with improved variation level set. This method enables two models organically combined to achieve the effect of mutual complementarities. Experimental results show that the proposed method has advantages of rapid, robust, and strong anti-noise and so on.

Key words: object extraction, variational level set, graph cuts, Graph Cuts Based Active Contours(GCBAC) algorithm

中图分类号: 

• TP391