Perceptual image hashing for DIBR 3D images based on ring partition and SIFT feature points

With a number of advantages, depth-image-based rendering (DIBR) has became an important technology in 3D displaying, as a result, more and more content-based image identification problems will turn out. Since either the center view with depth image or the synthesized virtual views could be illegally distributed, we need to not only protect the center views but also the synthesized virtual views with a novel method. In this paper, a novel perceptual hashing for DIBR 3D images is proposed, by dividing the center image into several rings, we select the suitable SIFT key-points in rings to form the final hashes sequence. Experimental results show that the proposed image hashing is robust to a wide range of distortions and attacks. Furthermore, it can ensure that the generated virtual images could be classified to the corresponding center image. When compared with the current state-of-the-art schemes, the proposed scheme can perform better identification performances under geometric attacks such as rotation attacks, and provide comparable performances under classical distortions such as additive noise, blurring, and compression.