Content-based image identification system with hierarchical scheme

In this work, we propose a content-based image identification system based on local descriptor and hierarchical indexing scheme obtained by clustering with global descriptor. The proposed system uses our local descriptor developed in order to achieve robustness against geometric distortion. Moreover, we take advantages of global descriptor by clustering original image database by edge histogram descriptor since the local descriptor is not guaranteed in matching time between images pairs. Our hierarchical searching scheme decreases computation time by fine matching within partial database of candidate images.