Based On3D Video Watermarking Algorithm Research

With the rapid development of network communication technology and digital media technology information dissemination and publishing becomes more and more convenient and simple, including the spread of the video image. For the original video content protection of information academia to discuss the hot and difficulty, and The Digital Video Watermarking technology is one of most important means copyright protection. In order to make effective to protect video copyright, we need to increase the robustness of the video watermarking and now people have become increasingly demanding on video HD, Therefore we must design strong robustness and invisible to protect the video copyright protection.For video watermarking, To some extent, it’s contradictory to ensure the strong robustness and invisibility in the video.So we will increase the embedding strength of the watermark information in order to prevent the video in the process of dissemination and released after the attacks can’t be extracted watermark information, so that it will cause to reduce the quality of video,but in pursuit of good quality of video, we have to reduce the watermark information embedding strength,this will result in video watermarking attacks after exaction is not be recognized, even can’t be extracted.above of both cases can’t achieve the purpose of protecting the video copyright.In this paper, the characteristics of3D video, video watermarking scheme based on the DC component of the DCT domain. Respective, the left and right view in3D video (or up and down view) to do the embedded watermark processing:8*8block DCT image transform, then collect stars2*2block with SVD transform, and use the DC coefficient Q(the quantization step) quantizing each piece is bigger corresponding singular values,to quantitative the results after embedding the watermark. Because DC coefficient SVD transform and SVD transform performs characteristic the meaning of the image rather than visual characteristics, so the program has a high level of robustness and invisibility.In addition,the program extraction method is blind extraction, the watermark are from the left and right view(or top and bottom view) will generating a plurality of watermark information.At last, we need determine the watermark information using vote type, watermark algorithm can resist median filtering, JPEG compression,noise and other attacks.