Variational color image segmentation via chromaticity-brightness decomposition

  • Zheng Bao*
  • , Yajun Liu
  • , Yaxin Peng
  • , Guixu Zhang
  • *Corresponding author for this work

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

7 Scopus citations

Abstract

A region-based variational model for color image segmentation is proposed using the chromaticity-brightness decomposition. By this decomposition, we extend the Wasserstein distance based method to color images. The chromaticity term of the proposed functional follows the data term of the color Chan-Vese model with constraint on unit sphere, and the brightness term is formulated by the Wasserstein distance between the computed probability density function in the local windows (e.g. 3 by 3 or 5 by 5 window) and its estimated counterparts in classified regions. Experimental results on synthetic and real color images show that the proposed method performs well for the segmentation of different image regions.

Original languageEnglish
Title of host publicationAdvances in Multimedia Modeling - 16th International Multimedia Modeling Conference, MMM 2010, Proceedings
Pages295-302
Number of pages8
DOIs
StatePublished - 2009
Event16th International Multimedia Modeling Conference on Advances in Multimedia Modeling, MMM 2010 - Chongqing, China
Duration: 6 Oct 20108 Oct 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5916 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Multimedia Modeling Conference on Advances in Multimedia Modeling, MMM 2010
Country/TerritoryChina
CityChongqing
Period6/10/108/10/10

Keywords

  • Chromaticity-brightness decomposition
  • Color image segmentation
  • Variational model

Fingerprint

Dive into the research topics of 'Variational color image segmentation via chromaticity-brightness decomposition'. Together they form a unique fingerprint.

Cite this