Abstract

As the internet becomes the basic resource of information, not only texts but images retrieval systems have been appeared. However, many of those supply only a list of images, so we have to seek the expecting images one by one. Although, image labeling is one of the solutions of such a problem, various words are labeled to an image if the words are extracted from only one Web page. Therefore, this paper proposes an image clustering system that labels images by words related to a search keyword. This relationships are measured by Web pages in WWW. By the experimental results, users were enabled to find the intended images more faster than the ordinal image search system.