Tracking a video object with the active contour (snake) predicted by the optical flow

This paper presents an application of the classical method of active contour (snakes) to the real time video environment, where tracking a video object is a primary task. The greedy method that minimizes the energy function to update the snake points, was combined with the optical flow method to cope with the situation that the object of interest moves too far beyond the reach of the edge sensor. The real time video processing platform has been developed to realize the deformable active contour by using VideoOCX of Visual Basic.