Identifying and Indexing Near-Duplicate Images Using Optimizing Technique in Web Search

Today's World Wide Web is growing drastically and duplicates occur in many fields. Importantly duplicate images that are uploaded into internet like a food product, document image, medical images, textile fields etc. So it becomes very important to identify those duplicate images. Near duplicates can be similar copies or differ a little in their visual content. Duplicate images introduce many problems of redundancy and copyright infringement in large set of image collections. This paper proposes a methodology for identifying and indexing the near duplicate images on web and optimizing the results. First step is to get the search image from the user and enhance the search image and then Features are extracted from search image using SURF (Speeded up Robust Features) that is to extract the local invariant features of search image. After this calculate the similarity measured among the features extracted images using sim-hash algorithm and then indexing Near duplicate images based on user’s search image using Locality Sensitive Hashing (LSH). And finally optimizing the results using Particle swarm optimization (PSO).We demonstrate that our identifying and indexing approach is highly effective for collections of up to a few hundred thousand images.