Fast snake algorithm for segmentation of cell image

An algorithm of boundary detecting automatically based on active contour model is proposed. The algorithm is applied to obtain accurate segmentation of one or more slices in a cell image. In order to get initial contour of snake model automatically, a synthetically method was introduced. First it got rough border by scale-independing algorithm of wavelet transform, then adopt morphological and B spline methods to get single pixel closed boundary as initial contour. Next.. to make active contour converged in actual boundary rapidly, a new exterior constrain force which correlates to image settled points was defined to replace the external energy of the traditional active contour model. The experiment shows that this algorithm can obtain the boundary of the desired object from cell images quickly and reliably with little user intervention. It also has practical value in the medical image analysis.