Robust Image Hashing Algorithm for Detecting and Localizing Image Tamper in Small Region

Paper proposes the novel technique for Active tamper detection and tamper localization, which uses image hashing algorithm applied on extended image based on ring partition and Non Negative Matrix Factorization (NMF). Key contribution is tamper detection in small region especially corner regions and exact localization of that tampered object. Particular tampered region is localized with the help of recursive use of image hashing algorithm on each partition of image in significantly less time. Experiments are carried out on various types of more than 200 images with different dimensions and memory space. Experimental results proved that this system has best perceptual robustness and discriminative capability. The system is able to detect and locate tamper at any region of image even from unrepresented corner region. The system is useful in many applications where original image plays an important role.