AMVP prediction algorithm for adaptive parallel improvement of HEVC

Author

Xiantao Jiang ; Tian Song ; Shimamoto, Takashi ; Lisheng Wang

Author_Institution

Dept. of Electron. & Inf. Eng., Univ. of Tongji, Shanghai, China

fYear

2014

fDate

17-20 Nov. 2014

Firstpage

511

Lastpage

514

Abstract

The Performance of High Efficiency Video Coding (HEVC) has been improved greatly compared with H.264/AVC including some parallel processing tools to achieve fast implementation on multi-core or many-core platforms. However, CU-level parallel processing is still a challenging issue. In this paper, the relation among spatial motion candidates is analysed firstly. Then, a parallelism spatial candidate list generation algorithm is proposed on the basis of the strong correlations among current prediction unit (PU), surrounding PUs and temporal correlated PUs. By evaluating the reliability of the motion vectors, the parallel granularity can be controlled adaptively. Experimental results show that the presented algorithm can achieve friendly parallel processing with only 0.26% BD-rate loss, based on the test model HM12.0.

Keywords

correlation methods; data compression; image motion analysis; multiprocessing systems; parallel processing; video coding; AMVP prediction algorithm; AVC; BD-rate loss; HEVC; HM12.0 test model; adaptive parallel improvement; high efficiency video coding; many-core platform; motion vector reliability; multicore platform; next generation video compression standard; parallel processing tools; parallelism spatial candidate list generation algorithm; prediction unit; spatial motion candidates; surrounding PU; temporal correlated PU; Bit rate; Encoding; Motion estimation; PSNR; Parallel processing; Prediction algorithms; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems (APCCAS), 2014 IEEE Asia Pacific Conference on

Conference_Location

Ishigaki

Type

conf

DOI

10.1109/APCCAS.2014.7032831

Filename

7032831