提出了一种自适应的Snake 算法,对于初始包络上的每一点,按照Greedy 算法中的能量公式计算其邻域内各点的能

量. 如果邻域不包含目标的轮廓,则将邻域半径加大,再次计算邻域内各点的能量,直到邻域内包含轮廓点为止. 该算法只需迭

代一次即可得到目标轮廓,而且增大了初始包络的收敛半径,并能够有效处理较高曲率凸形边缘的情况. 对比实验说明了这种

方法的有效性.-Snake proposed an adaptive algorithm, the initial envelope for each point on, in accordance with the Greedy algorithm in the energy formula of the neighborhood points of energy. If the neighborhood does not contain the target contour, the radius will be neighborhood increased, again the neighborhood of the points of energy, until the neighborhood of the contour points so far included. The algorithm only one iteration object contours can be obtained, but also increases the initial envelope radius of convergence, and can more effectively deal with convex edge of the case of high curvature. comparative experiments illustrate the effectiveness of this approach.