A Similarity Indexing Algorithm Based on Each Dimension Clustering

With the rapid development of multimedia information technology, multi dimensional indexing technique becomes an important research area with respect to the store and retrieval of visual information such as video and image To overcome the prominent ‘dimension curse’ problem, a novel method named similarity indexing algorithm is put forward based on each dimension clustering The method is used to realize each dimension clustering of the feature vector according to data distribution The process of realization can filter the irrelative data and reduce the search range step by step, hence speeding up the retrieval Experimental results show that the proposed method is suitable to perform the retrieval and search based on similarity indexing for the data with high dimension And it proves to be a simple and flexible indexing structure