Multiple-feature-based zero-watermarking for robust and discriminative copyright protection of DIBR 3D videos

Zero-watermarking is a key technique for achieving lossless and flexible copyright protection of depth image-based rendering (DIBR) videos. Existing approaches extract features both 2D frames maps via single mechanism to protect them simultaneously. However, it difficult these schemes fully satisfy the requirements two components, including remarkable discriminative capability 3D videos robustness against various attacks. Hence, in this paper, we propose novel multiple-feature-based zero-watermarking scheme DIBR To best our knowledge, first that integrates multiple improve Specifically, dual-tree complex wavelet transform discrete cosine enhance conversion noise addition, respectively, while ring-partition statistical residual ensure geometric attacks provide sufficient capacity. In use logistic-logistic chaotic system encrypt enhanced security design an attention-based fusion approach offer optimal solution. Extensive experimental results demonstrate proposed has stronger capacity compared state-of-the-art methods.