A dynamic high dimensional index scheme for contents based image retrieval

Abstract

High dimensional feature vectors are extracted for supporting the content-based image search. The indexing schemes for efficiently searching high-dimensional feature vectors are needed. In this paper, we propose a dynamic high dimensional indexing scheme for efficient contents based image retrieval. The proposed index scheme reduces the size of internal node information by eliminating internal node pointer. Also, it splits the internal node equally by using the overlapping split to maintain a height balanced tree. The performance evaluation results show that the proposed indexing scheme improves the response time compared with the existing schemes.