New Color Cluster Algorithm for Image Retrieval

To overcome the disadvantage of describing contents of an image only with global features, a new color cluster image retrieval algorithm based on object regions is proposed. In this paper, firstly, object regions are extracted from an image, and then a new color cluster algorithm is presented. The algorithm is developed based on HVS color space, histogram intersection and combination of two similarity values. The algorithm has been coded into a program, which runs under Windows system in a PC. It has been tested on a large image library, and the testing results show that it works fast and is irrelative to the image background colors.