A Review on Web-based Image Re-Ranking Techniques

Image re-ranking is a successful method for the web based image search. In the traditional techniques, the images are search based on only the query keyword given by the user and producing estimated output. But image search based on only query keyword is not efficient. Hence in image re-ranking, user intention is captured by clicking on one query image and retrieved the images related to query image. In this paper we have provided overview of traditional techniques such as Text-only related image search, Attribute based image search and image search using Query image. But these techniques face some limitations hence provide our proposed system which is Image search using semantic signature. Also the semantic signatures are short in size and specific to query keyword, hence this image re-ranking technique improves the accuracy and efficiency of image search.