A graph clustering algorithm with applications to content-based image retrieval

The graph is an important structure for representing objects and their relations. Its use in content-based image retrieval is still in its infancy, due to the lack of efficient algorithms for graph matching and graph clustering. Like (vector) data clustering, graph clustering plays a key role in organizing images according to their content. In this paper, we propose a new graph clustering algorithm based on the k-means algorithm. The key elements of the new algorithm are an efficient graph matching algorithm for computing the similarity between two graphs and a new median graph algorithm for computing the median of a set of graphs. Random graphs and a synthetic image database are used to show the performance of the proposed algorithm.