Texture discrimination with multidimensional distributions of signed gray-level differences

The statistics of gray-level differences have been successfully used in a n umber of texture analysis studies. In this paper we propose to use signed g ray-level differences and their multidimensional distributions for texture description. The present approach has important advantages compared to earl ier related approaches based on gray level cooccurrence matrices or histogr ams of absolute gray-level differences. Experiments with difficult texture classification and supervised texture segmentation problems show that our a pproach provides a very good and robust performance in comparison with the mainstream paradigms such as cooccurrence matrices, Gaussian Markov random fields, or Gabor filtering. (C) 2001 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.