Using association rules as texture features

A new type of texture feature based on association rules is proposed in thi s paper. Association rules have been used in applications such as market ba sket analysis to capture relationships present among items in large data se ts. It is shown that association rules can be adapted to capture frequently occurring local structures in images. Association rules capture both struc tural and statistical information, and automatically identifies the structu res that occur most frequently and relationships that have significant disc riminative power. Methods for classification and segmentation of textured i mages using association rules as texture features are described. Simulation results using images consisting of man made and natural textures show that association rule features perform well compared to other widely used textu re features. It is shown that association rule features can distinguish tex ture pairs with identical first, second, and third order statistics, and te xture pairs that are not easily discriminable visually.