Review Of Different Block Matching Motion Estimation Algorithms & Their Comparative Study

Abstract

Video compression is used to remove the temporal redundancy and  spatial redundancy by using video codec. Motion estimation (ME) is to  be done in the encoder side to find the best MV (motion vector) so that it  can be applied on stored frames by motion compensated block to  generate the predicted video in video compression. . In this paper, we  have taken various fast search techniques which are employed to find the  best matching point i.e. best motion vector (MV) in reference frame. In  this paper we have explained already implemented various block based  motion estimation techniques namely Full Search (FS), Three Step  Search (TSS), New Three Step Search(NTSS), Diamond Search  (DS),Four Step Search(4SS), Adaptive Rood Pattern Search(ARPS) . In  this paper, these six different BMA using motion estimation are  considered & implemented. Comparisons are made through the Peak  Signal to Noise Ratio (PSNR) and search points per macro block (i.e.  computation time) for different sizes of macro blocks and search areas.  The results predict that among all the algorithms, ARPS has the best  optimal PSNR based on computation time & algorithm.