Different methods of Web Image Re-Ranking
- Rahul C Kadam
- Prof. Nighot M.K.
Image search, image re-ranking, cluster of images, semantic space.
In Today’s world searching of images on internet are very popular, but most of the times searching result not exact match with the searching key. Improve the results of web-based image search as an effective way by Image re-ranking, has been adopted by current commercial search engines such as Google and Bing. Given a query keyword, pools of images are first retrieved based on textual information. When the user selects a query image from the pools of images, then the re-ranking of remaining images are based on their visual similarities with the user selected query image. A big challenge is that the similarities of images visual features do not well correlate with semantic meanings of images which interpret users search intention. Recently people proposed to matching images from semantic space which used reference classes or attributes closely related to the semantic meanings of images. Semantic signatures of images are improved both the efficiency and accuracy of image re-ranking. In this paper we discuss different methods for web image re-ranking and propose new re-ranking technique with removing of duplicate images.
Volume 4 Issue 3
Pages. 1063-1066