Study On Web Image Search Engine Based On MPEG-7

How to search needful image information from internet quickly and efficiently is an important research field. An archetypal system on Web Image Search Engine based on MPEG-7 is developed in this thesis.The traditional Content-Based Image Retrieval (CBIR) is no uniform standard of description on image content and no applying to Internet, which makes the system of CBIR not universal. Firstly, based on analysis about the standard of MPEG-7 (Multimedia Information Description Interface), an auto-generating method of XML based on MPEG-7 is proposed in this thesis, which provides a solution of universal description on the feature of image information. Secondly, based on the key technology of CBIR researching, a local CBIR module is implemented in this thesis, which constructs a general frame for the system of image search engine. Thirdly, through researching of technology about search engine, combing the technology of parsing image file and searching URL site, an Image Web Searcher is developed in this thesis ,when starts, it can provide image Data Source from Internet for the system. In the end, according to description of image feature based on MPEG-7, the frame of local CBIR and Image Web Searcher, under the framework of Struts, An archetypal system on Web Image Search Engine based on MPEG-7 is completed.This system is a content-based image search engine, which is accord with MPEG-7 standard. User can check out image information from Internet through inputting an example image. Testing result indicates that the system has attained the desired objective.