Wyner-Ziv Coding of stereo images with unsupervised learning of disparity

David Varodayan, Yao Chung Lin, Aditya Mavlankar, Markus Flierl, Bernd Girod

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

Abstract

Wyner-Ziv coding can exploit the similarity of stereo images without communication among the cameras. For good compression performance, the disparity among the images should be known at the decoder. Since the Wyner-Ziv encoder has access only to one image, the disparity must be inferred from the compressed bitstream. We develop an Expectation Maximization algorithm to perform unsupervised learning of disparity at the decoder. Our experiments with natural stereo images show that the unsupervised disparity learning algorithm outperforms a system which does no disparity compensation. It is also superior to conventional compression using JPEG.

Original languageEnglish (US)
Title of host publicationPCS 2007 - 26th Picture Coding Symposium
StatePublished - Dec 1 2007
Externally publishedYes
Event26th Picture Coding Symposium, PCS 2007 - Lisbon, Portugal
Duration: Nov 7 2007Nov 9 2007

Publication series

NamePCS 2007 - 26th Picture Coding Symposium

Conference

Conference26th Picture Coding Symposium, PCS 2007
CountryPortugal
CityLisbon
Period11/7/0711/9/07

    Fingerprint

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Varodayan, D., Lin, Y. C., Mavlankar, A., Flierl, M., & Girod, B. (2007). Wyner-Ziv Coding of stereo images with unsupervised learning of disparity. In PCS 2007 - 26th Picture Coding Symposium (PCS 2007 - 26th Picture Coding Symposium).