An adaptive and efficient selective multiple reference frames motion estimation for H.264 video coding

Yu Ming Lee, Yong Fu Wang, Jia Ren Wang, Yinyi Lin

Research output: Contribution to journalConference articlepeer-review

5 Scopus citations

Abstract

In the popular video coding standard H.264/AVC, many advanced techniques are employed. One important technique is the use of multiple reference frames motion estimation. However, the computational load increases with the number of references frames. In this paper, we suggest a selective multiple reference frames motion estimation (SMRFME) architecture which takes use of the information of the 1st reference frame to determine whether it is necessary to search remaining reference frames. In addition, three early termination schemes are applied to the remaining reference frames of the candidate modes. The simulation results demonstrate that the proposed algorithm can achieve up to 77% of time saving compared to the multiple reference frames full search algorithm, while maintaining a high coding performance.

Original languageEnglish
Pages (from-to)509-518
Number of pages10
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5414 LNCS
DOIs
StatePublished - 2009
Event3rd Pacific Rim Symposium on Image and Video Technology, PSIVT 2009 - Tokyo, Japan
Duration: 13 Jan 200916 Jan 2009

Keywords

  • AZB
  • Early termination
  • H.264
  • Multiple reference frames
  • Region based
  • Selective multiple reference frames motion estimation (SMRFME)

Fingerprint

Dive into the research topics of 'An adaptive and efficient selective multiple reference frames motion estimation for H.264 video coding'. Together they form a unique fingerprint.

Cite this