Robust and discriminative zero-watermark scheme based on invariant features and similarity-based retrieval to protect large-scale DIBR 3D videos

Digital rights management (DRM) of depth-image-based rendering (DIBR) 3D video is an emerging area research. Existing schemes for DIBR cause distortions, are vulnerable to severe signal and geometric attacks, cannot protect 2D frames depth maps independently, or have difficulty handling large-scale videos. To address these issues, a novel zero-watermark scheme based on invariant features similarity-based retrieval (RZW-SR) proposed in this study. In RZW-SR, extracted generate master ownership shares provide distortion-free, robust discriminative copyright identification under various attacks. Different from conventional schemes, our stores correlatively designs phase effective solutions addition, flexible mechanisms attention-based fusion designed maps, either independently simultaneously. The experimental results demonstrate that RZW-SR has superior DRM performance compared existing schemes. First, can obtain the relevant particular precisely reliably Second, ensures lossless, precise, reliable