Image Based Model for Document Search and Re-ranking

Rutuja G. Mhatre, Rajesh H. Bhise

Web Pages, search engines, multimedia search, ,document ranking.

Traditional Web search engines do not use the images in the web pages to search relevant documents for a given query. Instead, they are typically operated by computing a measure of agreement between the keywords provided by the user and only the text portion of each web page. This project describes whether the image content appearing in a Web page can be used to enhance the semantic description of Web page and accordingly improve the performance of a keyword-based search engine. A Web-scalable system is presented in such a way that exploits a pure text-based search engine that finds an initial set of candidate documents as per given query. Then, by using visual information extracted from the images contained in the pages, the candidate set will be re-ranked. The computational efficiency of traditional text-based search engines will be maintained by the resulting system with only a small additional storage cost that will be needed to predetermine the visual information.