Using color reach histogram for object search in colour and/or depth scene

  • Authors:
  • Naofumi Wada;Shun'ichi Kaneko;Tomoyuki Takeguchi

  • Affiliations:
  • Toshiba Co. Ltd., Tokyo, Japan;Department of Control and Information Engineering, Graduate School of Engineering, Hokkaido University, N13-W8 Sapporo, Japan;Toshiba Co. Ltd., Tokyo, Japan

  • Venue:
  • Pattern Recognition
  • Year:
  • 2006

Quantified Score

Hi-index 0.01

Visualization

Abstract

A method is proposed for object search in cluttered color and/or depth scene, which is based on color reach histogram. The histogram features an image by the distribution of pixel colors with their two- and three-dimensional locations. In real situation, it is sometimes difficult to keep illumination unchanged or stable, and then color distributions can address an another effective cue for object identification. The color reach histograms reveal strong robustness for illumination change and/or on-plane rotation. In comparison with conventional histogram-based methods, the stability and robustness in object search and tracking could be obtained and verified through many experiments with real objects.