Efficient texture feature extraction algorithm arbitrary-shaped regions

Region-based image retrieval has been an active research area. Texture is one of the important features to describe regions. In this paper, an efficient texture feature extraction algorithm for arbitrary-shaped regions is presented. With a proper initial padding used, the algorithm first extends an arbitrary-shaped region into a rectangular area onto which block transformation can be applied. Then a set of coefficients are selected, from which texture feature of the region can be extracted. Experimental results demonstrate the effectiveness of the proposed algorithm in texture feature extraction from arbitrary-shaped regions for image retrieval purposes. And it is shown that a proper initial padding is the key to the performance of the algorithm.