Application of Multidimensional Indexing Methods to Massive Processing of Multimedia Information

摘要

Since reduction of the processing cost is essential to large-scale multimedia applications, the processing tech- nique for database systems can play and important role in massive processing of multimedia information. This paper focuses on the effectiveness of two multidimensional in- dexing methods, the index construction method and the set-based nearest-neighbor search method. First, this paper presents a static construction method of the e Sphere/Rectan- gle-tree and evaluates its performance by applying it to the similarity retrieval of still images.