An efficient indexing method for similarity searches in high-dimensional image databases by principal axis analysis

This paper is focused on the design of the indexing structure for high-dimensional image databases by principal axis analysis. Firstly, we present the method of constructing the projection line with minimum inertia by principal axis analysis. Secondly, the proposed filtering mechanism, based on the projection scores of the database vectors, is presented. Then follows a method to enhance the discriminatory power of the approximations by incorporating the projection scores on multiple principal axes. Finally, some experimental tests to illustrate the effectiveness of our method are shown.