基于张量投票的主动轮廓边缘提取

Boundary Extraction of Active Contour Based on Tensor Voting

提出一种基于张量投票的主动轮廓边缘提取算法。该算法对图像进行张量编码,将其转化为二阶对称的半正定张量,每个输入张量通过预定的投票域对邻近数据进行稀疏投票,投票后每个输入点收集自身获得的选票,再进行稠密投票,以获得轮廓的显要特征,从而实现主动轮廓的边缘提取。实验结果表明,该算法能有效提取图像轮廓边缘,得到较好的物体主动轮廓模型。

This paper proposes an algorithm based on tensor voting to extract the boundary of active contour. It encodes the image and transforms it to a second order symmetric semi-definite tensor. Each input tensor takes spare voting to neighboring points through a predefined tensor field. After voting, the tensor collects acquired votes and takes dense voting for getting salient feature of contour. The boundary is extracted by implementing extremal algorithm for salient feature. Experimental results show that the algorithm can extract contour edge for image and get better active contour model of object.