Abstract

Local stereo algorithms are preferred for real-time applications due to their computational efficiency. Deciding the size of the required local support region is a challenging task. It fails to estimate accurate disparity for small support region and introduces fattening effect for big support region. Hence, a shape adaptive local support region is necessary to achieve accurate disparity. This paper proposes an anchor-diagonal-based shape adaptive support region construction for stereo matching. The proposed algorithm dynamically constructs local support region, and the aggregated matching cost is used for Normalized Cross-Correlation-based similarity measure. The algorithm is evaluated using benchmarked Middlebury stereo evaluation, and the obtained disparities are efficient as compared to state-of-the-art methods.

Original languageEnglish
Pages (from-to)893-901
Number of pages9
JournalSignal, Image and Video Processing
Volume9
Issue number4
DOIs
Publication statusPublished - 2015

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this