Analysis of Motion Estimation and Compensation for Video Compression: A Comparative Study of Motion Estimation Techniques for Video Compression

Abstract

This paper focuses on comparative study of the different motion estimation techniques for full search algorithms that have been proposed for inter frame coding in moving video sequences taking into consideration the different kinds of motion of the moving objects in successive frames. This is very important because it helps in bandwidth conservation by reducing temporal redundancy and reduction in power consumption by reducing computation complexity. It implements and compares 5 different types of block matching algorithms that range from the very basic Exhaustive Search to the recent fast adaptive algorithms like Adaptive Rood Pattern Search. The algorithms that are evaluated in this paper are widely accepted by the video compressing community and have been used in implementing various standards, ranging from MPEG1 / H.261 to MPEG4 / H.263. The paper also presents a very brief introduction to the entire flow of video compression. Keywords - Block matching, motion estimation, video compression, MPEG, H.261, H.263