基于极坐标中蛇模型的类圆目标轮廓提取技术研究

提出了转化到极坐标中的蛇模型.通过把蛇模型转化到极坐标中,使轮廓的候选点得以更有序的排列.由于采用了动态规划法并在整个能量空间中搜索能量泛函的极值,算法对能量泛函的局部极值有较强的鲁棒性.所提出的模型不需要确定初始轮廓,可以用非迭代方法直接求解.与传统的动态规划法和贪婪算法进行了比较实验.结果表明,所提出的算法对极坐标中极点的位置不是很敏感.

Extraction of the contours of elliptic objects based on polar-transformed snake

A polar-transformed snake for extraction of the contours of elliptic objects is presented. The candidate points for the contour are collocated by transforming the snake into polar coordinates. The proposed approach is robust to local minima of the energy functional because it is minimized by dynamic programming method in the whole energy space. A precise initial contour is not necessary and the solution can be obtained non-iteratively. The proposed algorithm is compared with the methods based on dynamic programming and greedy. Results show that the model is robust to local minima and the location of origin of the polar coordinates.