An Image Index Algorithm Based on Hierarchical Clustering

Along with the rapid development of multimedia technology and network technology, building an indexing structure in image information retrieval becomes a very challenging issue. In this paper, we analyze a cluster-based index approach-CLIMB, and also propose a novel idea of this index structure which called CBC-Tree. The clustering information is saved to the index file and also B+ tree is used to retrieve the last level of CBC-Tree. While combining the retrieval ability of B+ tree and CBC-Tree together, the efficiency of image retrieval is improved. Our experiment results show that the improved indexing structure is efficient.