Chromatic correlation features for texture recognition

  • Authors:
  • George Paschos

  • Affiliations:
  • -

  • Venue:
  • Pattern Recognition Letters
  • Year:
  • 1998

Quantified Score

Hi-index 0.10

Visualization

Abstract

The importance of chromatic correlations in representing chromato-structural images is investigated in this paper. A vision system consisting of two information channels, i.e., luminance and chrominance, is formed for the description of an image in both the xyY and HIS color spaces. Features from the correlation directional histograms are extracted. Chromatic correlations show a particularly robust behavior based on statistical redundancy tests. Classification results on a color texture set of 25 images in both the xyY and HIS color spaces show the discriminatory power of such features.