A robust method for interactive segmentation of Circle of Willis

The Circle of Willis (CoW) is a circulatory anastomosis that supplies blood to the brain and surrounding structure. Detecting and quantifying the CoW is essential for the prevention and treatment of aneurysm rupture and cerebral Infarction. In this paper, we present a robust interactive method to segment CoW mesh model from the whole cerebrovascular mesh base on graph-cut. First we construct the weighted graph G based on the input sketch. Then we find the minimum cost along the partitioned edges with Min-Cut/Max-Flow method. Moreover, we combine the shape-diameter function (SDF) and the angles between the normal vector of the neighbor face to construct the penalty energy which makes the method stable. Computational results show that the proposed interactive algorithm is robust, real time performance and friendly to user.