Contour Motion Estimation in Sequential Images

In this paper, we present a new method of velocity eld estimation for the points on the moving contours in a 2-D image sequence. The method decides the corresponding point in the next image frame by considering the curvature change of the given point on the contour. In the traditional methods, there are optical ow errors for the points which have narrow local curvature variations since those methods compute solutions by approximating normal optical ow. The proposed method computes the optical ow vectors of the contour points minimizing the curvature changes. Contours are extracted and tracked using snakes. We excluded the rearranging process and allowed the snaxel distance to vary, but the order of contour points is preserved. Each point on a contour has the unique corresponding point on the contour of next frame. The experimental results show that the new model computes accurate optical ow vectors for various moving contours .