Re-ranking of Images and Removing Duplicate Images

Image re-ranking is one of the most efficient way by which you can improve the image search results and has been adopted by many search engines. Texture analysis is popular operation in CBIR. In this paper we have combined text based search along with CBIR and removing of duplicate images is done using pixel matching algorithm. Text based search is done using tags. CBIR is done using texture feature. We have used haar wavelet decomposition for color images-mean clustering is used to form the cluster of images and similarity matching is done using F-norm theory. Experimental results reflect the importance of wavelets in CBIR. Our system improves the retrieval performance and better matches.