Efficient image segmentation for region-based motion estimation and compensation

An intra-frame segmentation strategy to assist region-based motion estimati on and compensation is presented. It is based on the multiresolution applic ation of a histogram clustering and a probabilistic relaxation-labeling alg orithm, followed by a local gradient-based bottom-up merging procedure. Spe cially suited for region-based video coding, it strongly differs from other proposals in that it generates arbitrary shaped image regions with pixel a ccuracy at a low computational cost, while allowing full reconstruction of the segmentation at the decoder without the transmission of any region desc ription information.