Image Retrieval Based on Color and Edge Direction Features
Zhang Cai-neng · Journal of Xiangnan University · 2006
In this paper,a new method of content-based image retrieval is presented,which integrates color and edge direction feature.First,we research the problems that existed in the conventional color histogram,and present a new color histogram;and then,we extract the shape feature of image by using edge direction histogram of non-segmentation,which is simple and effective,and performs retrieval independent of translation,scaling,and small viewing position variations.At last,similarity of two images is computed by two histogram features.In order to improve retrieval performance,we use a relevance feedback mechanism. This feedback enhances the retrieval effectiveness by dynamically modulating the weights of color-shape information.Experiments show that the proposed method is obviously superior to other methods.