This paper presents a new approach for visual tracking. We used deformable active contour model(i.e snake algorithm) as a powerful means for tracking a rigid object in coplaner motion. This active contour model is defined by an energy functional, and regarded as enregy minimization problem. To solve this problem, we used steepest descent method and brent's method. These numerical methods have fast convergent speed and high precision solution.
In this paper we summarize the strengths and weaknesses of the previous approaches and present a point algorithm which is modified for tracking of rigid object. In different of other approach, it take a serious view of behavior of snake in terms of force. It retains the improvements of stability, flexibility, and fast convergent speed. Assuming that the target moves in the rigid body motion, we constraint movement of snake.
The silhouette of rigid objects is captured by point algorithm. Using Green's theorem, the centroid is computed by means of the captured silhouette. Finally, the desired motion of the camera is computed with the objective of keeping target's position and shape invarient with respect to the camera frame.