Trademark image retrieval using hierarchical region feature description

摘要

A novel trademark image retrieval method is proposed in this paper. In this method, the image region of the trademark is iteratively partitioned into progressively smaller one along various directions and four region measurements are then conducted for features extraction against the partitioned regions. The merit of the resulting descriptors is that the geometrical features of different directions are finely captured in a hierarchical manner. A shifting feature matching scheme effectively guarantees the matching invariant to the rotation of trademark images. The experiments results demonstrate that the proposed method outperforms the state-of-the-art approaches for the trademark image retrieval.