Research Of Watermarking Based On Scalable Video Coding

The popularity of 3G networks, makes more and more people can use cell phones,MP4, laptops and other portable devices to browse video. The diversification of videoterminal requires the bitstream to implement scalability, JVT drafts an extension for H.264- Scalable Video Coding (SVC) in 2007. The appearance of SVC is very convenient forvideo service providers, also contribute to the further development of digital products,but cause the problems of copyright dispute growing. Digital watermarking technologywhich can solve copyright disputes and the content integrity and source authentication hasdeveloped from 90's last century. The research of Watermarking towards SVC has greatsignificance for the spread of SVC and the healthy development of the digital world.Scalable watermarking for image, by using the techniques in Computer Vision (CV)to extract the scale-invariant regions to embed watermark, but there exist some synchro-nized problems. For the video watermarking, there has been simple model for scalablewatermarking. This paper proposes a spatial watermarking algorithm which can protectcopyright based on the existed model of scalable watermarking and the standard of SVC,and introduces the concept of adaptive detection. Furthermore, an encryption algorithmand a watermarking algorithm can achieve content source authentication in compresseddomain also proposed based on the analysis of inter-layer prediction of SVC. These threecomponents can be integrated as a copyright protection system. The main organization ofthis paper is as follows:Firstly, proposes a spatial watermarking algorithm has scalability and high reliability.The analysis of watermark embedding space based on the spatial-layer technology ofSVC, which ensure watermark in the original video can be detected in every temporal,spatial and quality layer. Then an improved JND model for video in DWT domain isproposed, the watermarking algorithm based on this model can achieve good balancebetween the invisibility and robustness. The introduction of adaptive detection can avoidthe problem of more side information should be transmitted cause by multi-layer structureof SVC. Experimental results show that the proposed watermarking algorithm has goodperformance in the scalability of temporal, spatial and quality.Secondly, proposes a selective encryption algorithm and a watermarking algorithm to achieve content source authentication. The encrypted part and the watermark embed-ding space determined by the further analysis of the inter-layer prediction in SVC. Then,an easy to implement encryption algorithm in compressed domain can be used in spe-cific application scenarios is proposed, and a real time watermarking algorithm which haslower modified ratio is also proposed. Experimental results show that both algorithmshave good performance.