A Novel Approach for Image Representation and Matching Based on Mixed Graph Structure

A novel method based on mixed graph structure is proposed for image representation and matching. The mixed graph structure is constructed according to the spatial relation of the regions within an image. This structure does not require redundant information to describe images. An image matching process focuses on evaluating region attributes and relationships contained in the corresponding mixed graph structure. An association graph based on this mixed graph structure is then constructed. By searching the largest cliques in the association graph, we are able to obtain the optimal similarity among images under this structure. Experimental results are presented showing the effectiveness of our approach.