Spatio-temporal Filtering for Moving Objects Tracking

Motion eld segmentation is an important stage in a real-time object tracking. The multi-point optical ow approach is a means to achieve this, but the approach is limited by uncertainty. This paper presents a motion eld segmentation scheme which incorporates the multiresolution clustering and the multi-point method to overcome the uncertainty. This scheme produces a multireso-lution segmentation map, where the central region of a physical object is represented using coarse resolutions so as to achieve a better estimation of the optical ow, and the border region is represented using ne resolutions so as to achieve a better estimation of the object boundaries .