Comparison of Various Web Image Re - Ranking Techniques

Image re-ranking, is an quite an efficient way to improve the results that are fetched from the web-based image search query. Given a query keyword for the image, a pool of images are first retrieved based on textual information, then the images are re-ranked based on their visual similarities with the query image according to the user input. But when, the images� visual features do not match with the semantic meanings of the users� entered query or keyword, it becomes a major challenge to make available the actual searched image. Hence, in this paper, the various Web image Re- ranking techniques are studied, on how it approaches towards the Web Image search that the user has input in query.