A highly-parallel approach on motion estimation for high efficiency video coding (HEVC)

Author

Radicke, Stefan ; Hahn, Juergen ; Grecos, Christos ; Wang, Qijie

Author_Institution

Stuttgart Media Univ., Stuttgart, Germany

fYear

2014

fDate

10-13 Jan. 2014

Firstpage

187

Lastpage

188

Abstract

The new High Efficiency Video Coding (HEVC) standard will most likely be used in many applications in the near future. However, the encoding process is extremely complex and hence very time consuming. We propose a hybrid encoding architecture, which uses a Graphics Processing Unit (GPU) to perform uni-predictive Motion Estimation (ME) in a highly-parallel manner. This results in notable time savings of up to 57.41% for high definition video sequences, while the Rate Distortion (RD) performance is only insignificantly affected.

Keywords

graphics processing units; image sequences; motion estimation; parallel processing; video coding; GPU; HEVC standard; ME; RD; encoding process; graphics processing unit; high definition video sequences; high efficiency video coding; highly-parallel approach; hybrid encoding architecture; rate distortion performance; unipredictive motion estimation; Central Processing Unit; Computer architecture; Encoding; Graphics processing units; Motion estimation; Video coding;

fLanguage

English

Publisher

ieee

Conference_Titel

Consumer Electronics (ICCE), 2014 IEEE International Conference on

Conference_Location

Las Vegas, NV

ISSN

2158-3994

Print_ISBN

978-1-4799-1290-2

Type

conf

DOI

10.1109/ICCE.2014.6775965

Filename

6775965