A novel block matching based motion compensation using hybrid particle swarm optimization technique for efficient video compression

Fast and accurate motion compensation is a crucial task in all standard of video coding. Many block matching based motion compensation techniques are proposed to maintain good quality, less computational cost and good compression. But existing algorithms are not efficient in all respects. So, we need a motion estimation algorithm which will provide optimized output based on the above mentioned metrics. In this paper, we propose a new block matching based Particle swarm optimization technique for fast and accurate motion estimation and compensation. Spatial and temporal correlation is adapted for initial population. We followed some strategy for adaptive generations, particle population, particle location history preservation and exploitation. Experimental result shows our proposed algorithm gives the best or very near to best output in terms compression ratio, structural similarity index (SSIM) and peak signal to noise ratio (PSNR) and a huge improvement of 40%-72% in computational complexity in terms of average number of search points.