• DocumentCode
    2290746
  • Title

    Towards Fast and Robust Watermarking Scheme for H.264 Video

  • Author

    Gong, Xun ; Lu, Hai-Ming

  • Author_Institution
    Tsinghua Nat. Lab. for Inf. Sci. & Technol., Tsinghua Univ., Beijing
  • fYear
    2008
  • fDate
    15-17 Dec. 2008
  • Firstpage
    649
  • Lastpage
    653
  • Abstract
    Most available H.264 watermarking methods either require large expense of time or are rather fragile to intended attacks. In this paper, we present a novel watermarking scheme, which satisfies with the requirements of both rapidity and robustness. The proposed scheme directly embeds watermarks into H.264 bit streams at the decoder by modifying the quantized DC coefficients in luma residual blocks. To increase the robustness while maintaining the perceptual quality of the video, we employ a texture-masking-based perceptual model to adaptively choose the watermark strength for each block. A crucial problem of decoder-based watermarking algorithms is error drift, which refers to the watermark error accumulations among different blocks during intra or inter predictions. To eliminate the effects of error drift, we propose a drift compensation algorithm by estimating a compensation signal for each block before embedding the watermark bit. Experimental results show that the watermarks are highly imperceptible and resistant to common attacks. In addition, our scheme is faster and causes less increase in bit rate than previous methods.
  • Keywords
    prediction theory; video coding; watermarking; H.264 bit streams; H.264 video; compensation signal; decoder-based watermarking algorithms; inter predictions; intra predictions; perceptual video quality; robust watermarking scheme; Decoding; Discrete cosine transforms; Humans; Immune system; Information science; Information technology; Laboratories; Robustness; Video compression; Watermarking;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Multimedia, 2008. ISM 2008. Tenth IEEE International Symposium on
  • Conference_Location
    Berkeley, CA
  • Print_ISBN
    978-0-7695-3454-1
  • Electronic_ISBN
    978-0-7695-3454-1
  • Type

    conf

  • DOI
    10.1109/ISM.2008.16
  • Filename
    4741242