Texture recognition or classification using statistics

Texture recognition/classification is one of the long standing problems of image processing. In this paper we regard this task as a 'pattern recognition problem in the statistical parameter space'. There are three stages in this process - statistical feature extraction, statistical feature selection, and classification. We focus on the second and third order statistics for feature extraction. We use a number of criteria for feature selection and a number of techniques for classification. Results from the experiments are included.