Research On Video Watermarking Scheme Based On Local Pseudo-Zernike Moments

Digital watermarking technology embedding some of the information (that isdigital watermarking) directly into digital carriers (including multimedia, documents,software, etc), does not affect the carrier's original value, and the information can noteasily be removed. In assurance cannot be seen, digital watermarking can be extractedwith special watermark algorithm after attacks. This thesis systematically summarizesthe significance,key techniques,main features and the research status of videowatermarking technology, focuses on video watermarking algorithms and principles,and analyzes its strengths and weaknesses. Finally, this thesis introduces theapplications of image moments theoryin the watermarking technology.Tiny geometric attacks (rotation,scaling,translation) will be able to make theimages or video watermark destruction, led to the failure of the watermark detection. Asa result, the design of the resistance to geometric attacks video watermarking algorithmbecome a new research direction of digital watermarking technology. For rotation,scaling and translation(RST) attacks can desynchronize the watermark detection so thatmany watermark systems fail, this thesis presents a new video watermarking schemebased on local Pseudo-Zernike moments. Using the RST invariant property ofPseudo-Zernike moment magnitudes of a normalized image, the watermark image isembedded into the centre circle area of I frame in the spatial domain after reconstruction.As high similarity between video frames, an optimal moment set is selected accordingto I frame and the adjacent P frame. In detection, the Pseudo-Zernike moments of thepossibly attacked I frame are computed first and the Pseudo-Zernike moments oforiginal I frame are substituted by the Pseudo-Zernike moments of the adjacent P frameselected in the optimal moment set as the key to extract watermark. Simulationexperimental results show that the video watermarking algorithm based on localPseudo-Zernike moments can get good quality images, improve the algorithm'sdependence on original data, and makes the watermark more robustness to geometricattacks(arbitrary angle of rotation attacks and scaling attacks)than the global moments atthe same time.