An optimally fast greedy algorithm for active contours

Greedy algorithm is an iterative method used for minimizing the energy functional of active contours, snakes. In this paper we will show that using a suitable set of search patterns instead of a full search can increase its speed. A search pattern set that expands the search space and minimizes a suggested computational complexity cost function achieves the maximum speed. Finally a comparative study is made to show the agreement of the theoretical derivation with experimental results.