An improvement algorithm of the snake model

The traditional snake model often can not resolve the contradiction between algorithm complexity and searching the global optimization. A new tracking algorithm is proposed to optimize the snake mode in this paper. First, using the cg of the snake initial contour, the new algorithm draw a set of rays which intersect with the contour, and it keeps the snake testing points and their searching range on the rays. It simplifies the searching dimension from 2D to 1D. And then in the process of the snake points moving along the rays, the acceptance of the snake points is determined by the simulated annealing algorithm. it makes the snake model reach global optimization with the increase of the iteration. Applying the algorithm to the segmentation of leucocytes, the segment speed is faster and the accuracy is higher compare to the original algorithm. The result shows that the new algorithm can satisfy the real-time request and has a strong practical and research value.