Motion compensation with higher order motion models for HEVC
Cordula Heithausen, Jan Hendrik Vorwerk · 2015
In recent video compression standards only translational motion is accurately compensated, constraining any higher order motion to be approximated by being split into smaller translational units. The objective of this paper is to improve the coding efficiency for video sequences containing complex motion.Various higher order motion models are considered and evaluated to this end. Motivated by the higher maximum block size introduced by the High Efficiency Video Coding Standard (HEVC) and a suitable combination of motion parameter estimation, interpolation and coding, a higher order motion compensation framework for the HEVC is proposed. Through this an average data reduction of 2.9% as well as an increase in average block size is achieved.