ENHANCED WEB IMAGE RE-RANKING USING SEMANTIC SIGNATURES

Several commercial search engines adopt Image re- ranking approach to enhance the quality of results for web based image search .When an user issues a query keyword, the search engine first selects a group of images based on textual information. If the user selects a query image from the retrieved group, the rest of the images are re-ranked based on their visual similarities with the query image. The images with similar visual features cannot be correlated. Also learning a universal visual semantic space which depicts the characteristics of images which are highly different to each other is a tedious task.