Compressed Domain Video Watermarking Technology Based On H.264

Information society is promoted by the rapid development of communicationtechnology, wireless network and computer technology. The digitization andnetworking of information cross the limitation of space and time, as the same time,the development of video information industry is influenced. The problem of videocopyright and certification is implicated in the growth of users’ demand and theapplication of new media with the video content is enriched. Therefore, how toguarantee the safety of video content has been an important foundation of existenceand development of video information industry.Digital watermarking technique is one of the efficient methods to solve theproblem of video copyright. Video watermarking algorithm combined with thecompress coding standard has a broad applied prospects, since the video data isvery large and is limited by transmission bandwidth, the requirements for storageand transmission are relatively high. By combining with H.264video compressioncoding standard, an initial research is made for the digital watermarkingpreprocessing scheme and the compress video watermarking algorithm, the mainwork and innovations of this dissertation are as follows:(1) The elementary theory about video watermarking and H.264videocompress coding standard are studied. First, the development history andtechnology features of H.264standard are summarized and the principle of theencoder and decoder of H.264standard is also described. Second, the videowatermarking technology is introduced in detail, including the system models andcharacteristics of digital watermarking and the common attacking methods.(2) A watermark preprocessing scheme combined with Arnold transform andsecret sharing is proposed. For the low security of Arnold transform and thedrawbacks of key space, the ideology of secret sharing is introduced into thepreprocessing scheme. This scheme is able to achieve effective compression andenhance the confidentiality of the watermark information. The watermarkpreprocessing scheme proposed in this dissertation is universal which can be usedby different carriers and a more perfect system about copyright protection andauthentication can be built by combing with different watermarking algorithms.(3) An adaptive video watermarking scheme based on DCT coefficients is proposed. The medium frequency coefficients of DCT is selected when thewatermark information is embedded in this scheme, the revised extent of quantizedcoefficients is controlled by the average energy of current block, therefore, thescheme proposed here has the advantages that embedded speed is quick anddistortion is small. The test results show that this algorithm is able to resist therequantization attack and noise attack on the premise that the video quality isguaranteed. Meanwhile, the extracted watermark information after different attacksstill be able to meet the requirements of video certification. The imperceptibility ofwatermark information and the robustness of conventional attack are both takeninto account in this video watermarking algorithm.