A Fast Macroblock Mode Decision Algorithm for MVC Based on SVM

Author

Huan Dou ; Zhipin Deng ; Kebin Jia

Author_Institution

Dept. of Electron. Inf. & Control Eng., Beijing Univ. of Technol., Beijing, China

fYear

2012

fDate

18-20 July 2012

Firstpage

387

Lastpage

390

Abstract

This paper presents a fast macroblock mode decision algorithm based on Supporting Vector Machine (SVM) for MVC on the JMVC platform. The corresponding coded information of the reference view is used to design a three-stage SVM classifier for the proposed mode decision algorithm. Experimental results indicate that the proposed algorithm can save remarkable encoding time while maintaining nearly the same visual quality of the original sequences.

Keywords

image classification; image sequences; support vector machines; video coding; H.264/AVC; JMVC platform; coded information; encoding time; macroblock mode decision algorithm; multiview video coding; supporting vector machine; three-stage SVM classifier; video sequence; visual quality; Algorithm design and analysis; Classification algorithms; Encoding; Signal processing algorithms; Support vector machine classification; Video coding; H.264/AVC; Mode decision; Multiview Video Coding (MVC); Supporting Vector Machine (SVM);

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP), 2012 Eighth International Conference on

Conference_Location

Piraeus

Print_ISBN

978-1-4673-1741-2

Type

conf

DOI

10.1109/IIH-MSP.2012.100

Filename

6274261