Web Image Re-ranking using Semantic Signature and Visual Query Expansion

Abstract

Web mining is a important aspect for users to get accurate and relevant data as per user need and satisfaction.Web image reranking is an effective way to improve the results of web based image searching.A major challenge in web image reranking is that the similarities of visual features are not correlated with images and semantic meanings which can not interpret users search intention.In proposed approach, By providing keyword expansion to textual query and visual query expansion to image query the results retrieved are promising as compared to existing system.Feature extraction, Hashing, Euclidean distance are used to make semantic signature more compact and to further enhance efficiency.