A fast image retrieval using the unification search method of binary classification and dimensionality condensation of feature vectors

Abstract

We present the two-stage content-based image retrieval as a new fast image retrieval approach using the unification search method of binary classification and dimensionality condensation of feature vectors. The method successfully reduces the overall retrieval time, while maintaining the same retrieval relevance as the conventional exhaustive search method. By the extensive computer simulations, we have observed that the method is more effective as user-specific threshold for the similarity score increase. © Springer-Verlag Berlin Heidelberg 2005.