Searching of near exact duplicate images in cloud database

Abstract

Image modification becomes very easy because of image editing software and rapid development of Smartphone. Image is a way of communication, on social site people post their picture and other people either forward it or perform some operation like zoomed in/out, rotation and repost their image which is not identical image, it is near-exact duplicate image which resulted in large amount of images so it increase the storage cost. In this we use binary local image descriptors BRISK, and binary hashing methods, like as Locality Sensitive Hashing (LSH). The purpose of paper is to retrieve the near-exact duplicate image of the query image. For feature detection and extraction we use BRISK and LSH is used for indexing purpose which maps the similar image in the same bucket and use hamming distance for checking the similar images.