Object category retrieval for multimedia databases

In this paper, a novel content-based object category retrieval technique using salient regions and moment invariants is presented. The salient regions are detected according to Shannon entropy and scale selection. They are invariant under various viewpoint transformations, and intra-class variations. These regions are the most informative ones, hence are potentially more effective for image indexing and retrieval. Generalized colour moment invariants are used as the invariant descriptor on the selected salient regions to characterize them. The distance of the moment invariant vectors calculated from regions of different images is evaluated to be the similarity measure between images. The experimental results show that our proposed method is very efficient and effective on retrieving object categories from image databases.