A fast method to determine co-occurrence texture features

Abstract

A critical shortcoming of determining texture features de- rived from grey-level co-occurrence matrices (GLCM's) is the excessive computational burden. This paper describes the implementation of a linked-list algorithm to determine co-occurrence texture features far more efficiently. Behavior of common co-occurrence texture features across difference grey-level quantizations is investigated. I. INTRODUCTION Texture features calculated from grey-level co-occurrence matrices (GLCM's) are often used for remote-sensing image interpretation (1)-(3). There are acknowledged computational shortcomings when using GLCM's to determine co-occurrence texture features. Such re- strictions make pixel-by-pixel image segmentation impractical. Shokr (3) suggested that a linked list approach may be better suited for generating co-occurrence features than a matrix approach. The focus of this communications is to describe such an implementation and to provide insight into its performance.