The Research And Application Of Video Watermarking

As the fast development of network and multimedia technology, the application range of digital technology and products has arisen. And the dissemination, production and acquiring of video media has been easy, at the same time the numerous violations of illegal copying, distribution and tampering have appeared, which violate the interests and legitimate rights of producers and consumers, therefore the content protection and copyright issues of digital products aroused extensive attention. Then digital video watermarking emerged. The digital watermarking can protect the integrity of digital product content and the ownership by embedding specific data to original data. And it is widely used in the field of Medical science, legal evidence, notes security, etc. Digital watermark has become a focus in the research area of digital media. The research and application of digital watermark has very important theoretical sense and application value in the aspect of ensuring the information of copyright, data authentication and integrity test, testing integrity and authenticity of data, controlling of radio broadcast, and so on.This thesis discussed the foundational technology issues of watermarking, described characteristics, the research background, application area, research status, classification, and development trends of video watermark. In the meanwhile researched several algorithms that can be used in video watermark. Designs and implements Video Watermarking Algorithm Based on SURF Video for Segmentation. Main results are summarized as follows:(1) This thesis presents a video splition algorithm based on SURF feature points, deeply study SURF local gray feature, extract SURF feature points and design a segmentation according to the distribution of SURF feature points of video frames. The algorithm detects the shot boundary by analyzing the nature that the matching number of SURF feature points of frames change over time. After obtaining the reliable matching point by using the above method, the lens can be detected. In general content within a shot scenes change little, which change significant at the switching shot. The nature can be well reflect through the changing of SURF matching feature points between the front and back frames. Thus, shot boundary can be accurately detected by analyzing the process of SURF matching feature points changing over time in the video. As with the SURF feature points of scale invariant, noise invariance, rotation invariance, this method can accurately split lens, under the processing of video and the general attack on the video content.(2) This thesis presents a singular value decomposition video watermarking algorithm based on ICA frames, analysis various kinds of video watermark algorithm that resisting synchronization attack on space and time, collusion attack. Towards most existing video watermark technology can not resist these types of attacks, this mothed improves robust of video watermark for video synchronization attack in the time, by locating the video content through video feature information, enhances the robust of programs for video internal collusion attack by embedding watermark to the principal, raises the robust of watermark for synchronization attack in the space by embedding watermark algorithm to singular value of ICA frames. Experiments show that, the algorithm of this thesis has better robust to most of attack, especially improved the robust that resists the simultaneous attacks in the time and space, collusion attacks. The method of block svd fully uses three characteristics of singular value, has greatly invisibility, and achieves the binding detection.