A new adaptive return prediction search algorithm for block matching

Chih Lun Fang, Wen Yen Chen, Yuan Chen Liu, Tsung Han Tsai

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

In most block-based video coding systems, some of the fast block matching algorithms use the origin as the initial search center. But the tracking result of these algorithms is not well enough. To improve the accuracy of the fast block matching algorithms, a new adaptive return prediction (ARP) search algorithm is proposed in this paper. The proposed algorithm exploits the temporal correlation and the characteristic of returning phenomenon to obtain one or two predictive motion vectors, and one of the predictive motion vectors are chosen as the initial search center. This predicted search center is found closer to the global minimum, thus the center-biased algorithm can find the motion vector more efficiently. Simulation results show that the proposed algorithm enhances the accuracy of BMA as well as reduces their computational requirement.

Original languageEnglish
Title of host publicationAdvances in Multimedia Information Processing - PCM 2002 - 3rd IEEE Pacific Rim Conference on Multimedia, Proceedings
EditorsYung-Chang Chen, Long-Wen Chang, Chiou-Ting Hsu
PublisherSpringer Verlag
Pages120-126
Number of pages7
ISBN (Print)3540002626, 9783540002628
DOIs
StatePublished - 2002
Event3rd IEEE Pacific Rim Conference on Multimedia, PCM 2002 - Hsinchu, Taiwan
Duration: 16 Dec 200218 Dec 2002

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2532
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd IEEE Pacific Rim Conference on Multimedia, PCM 2002
Country/TerritoryTaiwan
CityHsinchu
Period16/12/0218/12/02

Fingerprint

Dive into the research topics of 'A new adaptive return prediction search algorithm for block matching'. Together they form a unique fingerprint.

Cite this