COMBINING CLUSTERING AND RELEVANCE FEEDBACK TECHNIQUES IN CONTENT-BASED IMAGE RETRIEVAL

摘要

This paper presents a summary of results of ongoing research on the impact of combining clustering and relevance feedback techniques in a content-based image retrieval (CBIR) system. It describes a novel approach to learning the user's goals and the relevance of each feature for a particular search, by employing a combination of Bayesian learning algorithm and cluster analysis techniques. Results of experiments show that the proposed model exhibits good performance for moderate-size, unconstrained databases, encouraging further research and improvements.