Skip to main navigation Skip to search Skip to main content

Stereo matching with improved radiometric invariant matching cost and disparity refinement

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

Abstract

Accurate and real-time stereo correspondence is a pressing need for many computer vision applications. In this paper, an improved radiometric invariant matching cost algorithm is proposed. It effectively combines modified census transform with relative gradients measures. Although it is very simple, comparison results on Middlebury stereo testbed demonstrate that it has much lower error rates than many existing algorithms and is very close to the ANCC algorithm which represents the current state of the art under extreme luminance condition but outperforms the ANCC algorithm greatly when there are small radiometric distortions. In addition, we also develop a disparity refinement method with computational complexity invariant to the disparity range. Experimental results on Middlebury datasets show those artifacts near object boundaries are reduced using the proposed disparity refinement method.

Original languageEnglish
Title of host publicationIntelligent Computing - 12th International Conference, ICIC 2016, Proceedings
EditorsDe-Shuang Huang, Vitoantonio Bevilacqua, Prashan Premaratne
PublisherSpringer Verlag
Pages61-73
Number of pages13
ISBN (Print)9783319422909
DOIs
StatePublished - 2016
Event12th International Conference on Intelligent Computing, ICIC 2016 - Lanzhou, China
Duration: 2 Aug 20165 Aug 2016

Publication series

NameLecture Notes in Computer Science
Volume9771
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference12th International Conference on Intelligent Computing, ICIC 2016
Country/TerritoryChina
CityLanzhou
Period2/08/165/08/16

Keywords

  • Census transform
  • Disparity refinement
  • Radiometric invariant
  • Stereo matching

Fingerprint

Dive into the research topics of 'Stereo matching with improved radiometric invariant matching cost and disparity refinement'. Together they form a unique fingerprint.

Cite this