With the increased usage of image databases over World Wide Web for forensic, medical and other applications, researchers focus on improving the efficiency of retrieving the most relevant images for a query image. This becomes possible by refining the retrieval process based on the user feedback. The proposed work extracts Tchebichef moments and retrieval rate is enhanced by employing interactive Genetic Algorithm. Since Tchebichef moments are orthogonal and hence non redundant, the retrieval rate is promising. Experiments were carried out on COREL images and the average retrieval rate of 75% reveals the efficacy of the proposed work.

Keywords. Moment features, Image retrieval, genetic algorithm