Motion estimation Using Color Based Segmentation

In block based motion estimation, each frame is divided into macroblocks and motion estimation is performed macroblock wise. It has high computational cost. In this paper, a block based motion estimation technique using color segmentation has been proposed which has single block for each object and motion estimation based on diamond search technique has been used to compute motion vectors. Experimental results show that the performance of Diamond search technique using color segmentation is better than conventional Diamond Search Technique in terms of PSNR, average number of search points and compression ratio.