Precise image retrieval on the web with a clustering and results optimization

Effective image searching in WWW has become important to various users, and the image meta-search engine is an effective technique to improve the quality of retrieval results of Web images on the Internet. The emphasis of the thesis is to propose a model of image meta-search engines, and a vectorization method was adopted to apply HACM (hierarchical agglomerative clustering methods) clustering techniques on images search that are then optimized by a specially designed genetic algorithm. The method provides a more significant and restricted set of images as the final result for a user´s search on an image meta-search engine. Some experiments have been run on to test the image meta-search engine. The method enables the image meta-search engine to handle a query term in a reasonably short time and return the results with high accuracy.