6 Image Similarity Measurement using Shape Feature

In this paper, we describe an incipient method for image retrieval predicated on the local invariant shape feature, designated scalable shape context. The feature utilizes the Harris-Laplace corner to locat the fix points and coinside scale in the animal and flower image. Then, we utilize shape context to explain the local shape. Correspondence of feature points is achieved by a weighted bipartite graph matching algorithm and the homogeneous attribute between the query and the indexing image is presented by the match cost. The practical results show that our method is efficient than shape context and SIFT for the animal and flower image retrieval