Shape matching and retrieval using vector functions

In this paper, we propose a shape matching and retrieval method using vector functions and selective perceptually motivated morphological strategy. For each sample point, we extract a set of vectors originating from the point to all other sample points on a shape, and the set of vectors in order is called a vector function. During the definition of matching cost between two sample points, two components of the vectors are used. In addition, selective morphological strategy is used to improve vector functions, in which a simple and efficient algorithm to measure morphological process necessity is provided. With a threshold, the vast majority of time consuming matching between morphological processed shapes could be avoided. Our experiments on MPEG-7 database indicate that our method outperforms the state-of-the-art shape matching algorithms.