Re-ranking of images is an effective way to improve the results of web-based image search and it has been adopted by the current search engines. When a query keyword is given, a list of images isretrieved first on the basis of textual information given by the user.Nowadays, with user's search intention, web search engine gives a set of relevant as well as irrelevant Images and then the user selects query image from the pool. The remaining Images are reranked based on their visual similarities.

more »

... he main drawback with current web search engine is that visual similarities of images do not correlate well with Images semantic meaning. We propose a unique web Image re-ranking framework that provides learned Images visual and semantic meaning based on query keywords.