Application of Bi-partite Optimal Matching in Color-Base Image Retrieval

Image description and characteristic extraction are the most important steps of color-Base image retrieval. This paper proposed a new method based on the related methods of graph theory. The new method defines every parts of the images as a vertex, than calculate the similarity measure as the weight value to design a weighted bi-partite graph. The congruent relationship between images is transform into searching the optimal matching in this bipartite graph. In the new method, both color dimensional distribution and geometric transformation invariance are considered. The simulation results show the effectiveness of the novel method.