Inter-frame and inter-layer motion prediction during fast block motion estimation in MCTF

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The motion compensated temporal filtering (MCTF) using 5/3 LeGall bi orthogonal wavelet filter requires bidirectional motion estimation (ME) that is computationally intensive. Hence encoder of Scalable Video Coder(SVC) becomes very slow. This paper proposes Motion Prediction technique for Fast Block Based ME during MCTF in SVC. The technique exploits center biased characteristics of motion vectors. It applies coarse and fine search of any fast block based ME, only to the first pair of frames in a group of pictures (GOP). The generated motion vectors (MVs) are supplied to the next consecutive frames and even to subsequent temporal levels, that is used to initiate the fine search. The technique significantly reduces the number of blocks that undergoes ME in a GOP and hence reduces computational complexity of ME in MCTF. The proposed algorithm is implemented in MC-EZBC and there is no visible variations in the quality of the decoded video.

Original languageEnglish
Title of host publicationVisual Information Systems
Subtitle of host publicationWeb-Based Visual Information Search and Management - 10th International Conference, VISUAL 2008, Proceedings
Pages150-154
Number of pages5
Volume5188 LNCS
DOIs
Publication statusPublished - 2008
Event10th International Conference on Visual Information Systems, VISUAL 2008 - Salerno, Italy
Duration: 11-09-200812-09-2008

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5188 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Visual Information Systems, VISUAL 2008
Country/TerritoryItaly
CitySalerno
Period11-09-0812-09-08

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Theoretical Computer Science

Fingerprint

Dive into the research topics of 'Inter-frame and inter-layer motion prediction during fast block motion estimation in MCTF'. Together they form a unique fingerprint.

Cite this