Extraction of video object based on object tracking and matching

This paper presents a new approach to the segmentation of video object plane automatically. The method is to extract the initial binary contour models of moving object from several frams by using morphological motion filter. Then an object tracker is used that matches a two-dimension binary model of the obect agaisnst subsequent frames using the Hausdorff distance. In order to allow for the change of shape of object the object twcher based on active contour model is presented to match the object in the new location. From a series of binary contours we can extract the moving object in the end. Experiments show that our algorithm can extract video objects and video object planes from sequences.