A Photo Identification Framework to Prevent Copyright Infringement with Manipulations

In recent years, copyright infringement has been one of the most serious problems that hamper development culture and arts industry. Due to limitations existing image search services, these infringements have not properly identified number increasing continuously. To uncover handle big data extracted from photos, we propose a photo identification framework accurately manipulations stolen photos. From collage cropped regions interest (RoIs) are detected reduce influence cropping identify each by Image RoI Detection. Binary descriptors for quick database generated RoIs Hashing robustly geometric color manipulations. The matching results verified measuring their similarity using proposed Verification false positives. Experimental demonstrate outperforms other retrieval methods in accuracy significantly reduces positive rate 2.8%. This is expected practical situations effect on market.