ENHANCED PREDICTIVE MOTION VECTOR FIELD ADAPTIVE SEARCH TECHNIQUE (E-PMVFAST) ¡V BASED ON FUTURE MV PREDICTION (FriAmOR3)

Author(s) : Hoi-Ming Wong (Hong Kong University of Science and Technology, Hong Kong) Oscar C. Au (Hong Kong University of Science and Technology, Hong Kong) Chi-Wang Ho (Hong Kong University of Science and Technology, Hong Kong) Shu-Kei Yip (Hong Kong University of Science and Technology, Hong Kong)

Abstract : Motion Estimation (ME) is a core part of most modern video coding standard, and it directly affects the compression efficiency and visual quality of a video. If Full Search (FS) algorithm is used, ME could takes over 70% of computational power. Many algorithms such as TSS and PMVFAST, have been developed to achieve great speed up for ME. In this paper, a new algorithm Enhanced-PMVFAST (E-PMVFAST) is proposed, which performs better than most if not all other existing algorithms in terms of speed up factor while keeping the similar PSNR with FS. Our experiments have also verified the robustness of the proposed E-PMVFAST algorithm.

Menu