Motif matrix inference for rotated image indexing and retrieval

摘要

Motif is a promising descriptor to depict the content of image. In this study, two motif-relevant matrices, i.e. a motif average matrix (MAM) and a motif excessive matrix (MEM), are proposed firstly to describe the color and texture features of an image. Subsequently, in the light of the inference of MAM and MEM, a motif matrix (MM) is further proposed to resolve the issues of rotated image retrieval. In terms of such an inference MM, a 256 * 8 matrix, incorporates the colorful and textural characters and represents the consistent feature between the original and its rotated images. That is, MM reveals the potential relevance for rotated image retrieval. We carry out the experiments on the benchmark Corel image dataset, and the experimental results show that our approach of motif matrix inference improves the retrieval performance in comparison with the state-of-the-art image retrieval approaches.