Sub-Pixel Motion Estimation using Wavelet Based Counterlet Transform

Conference: International Conference on Signal, Image Processing Communication and Automation
Author(s): Vidyasagar K N, Bharathi S.H Year: 2017
Grenze ID: 02.ICSIPCA.2017.1.35 Page: 218-224

Abstract

Better Compression of Video Frame can be achieved by high accuracy estimation and compensation of motion vectors to diminish the estimated Error between P and I frame. Compensating the motion in a video by Quarter pixel has been accurate, on the other hand coding rate of the encoder will be very slow for block matching Sub-pixel exhaustive Search, targeting this a novel Sub-pixel motion estimation procedure using wavelet-based counterlet transform (WBCT) is presented in this paper. Primarily, the optimum low occurrence subband whole number pixel point of the video frame block in transform domain is found out. Then by considering the optimum pixel point as a middle point to search the optimum motion vector of the subpixel using chaotic particle swarm optimization (PSO) algorithm. This optimum motion vector is then considered as initial motion vector of high occurrence sub-band in video frame block. To finish with, the high occurrence sub-band is subjected for thorough search to get optimum motion vector. As the experimental results show, the proposed methodology has better performance for search time and motion compensation effect in standard video sequences.

<< BACK

ICSIPCA - 2017