Abstract:

Highdimensional indexing technique, as the quick query strategy in high dimension data, has been widely used in ContentBased Image Retrieval. This paper builts a highdimensional indexing structure to store the eigenvalue of image inside using Rtree which is firstly proposed by Guttman in 1984. In order to improve the efficiency of image retrieval, this paper first introduces the structure of Rtree. Then the advantage of Rtree searching is analyzed by comparing the query times and searching time of Rtree searching and linear structure searching respectively. The experimental results demonstrate that, benefited from the Rtree structure, the query times and searching time are reduced and the efficiency are improved remarkably.

Key words: highdimensional retrieval, Rtree, image retrieval, spatial data, socket