A Potential Heuristic-based Block Matching Algorithms for Motion Estimation in Video Compression

Abstract

Motion estimation (ME) is one of the element keys in video compression that takes up to 60% in processing time. Block matching algorithm (BMA) is a technique that is used to reduce the computational complexity of ME algorithm due to its efficiency and good performance. Strategy of searching is one of the factors in developing motion estimation algorithm that has the potential to provide good performance. This study aims to implement several selected BMAs for achieving the least number of computations and to give better Peak Signal to Noise Ratio (PSNR) values using different video sequences. The proposed algorithms are modified based on the search strategy adapted from the standard algorithms approach. The results have proved that both modification algorithms (MDS and MARPS) have the potential in reducing the number of computations and achieved good PSNR values in all motion types as compared to DS and ARPS respectively. This work could be improved by using metaheuristic algorithms approach such as particle swarm optimization (PSO), artificial bee colony (ABC), tabu search (TS) and etc to provide the better result of PSNR values without increasing the number of computations