A semi-fragile content authentication algorithm is proposed for low bit-rate H.264/AVC video in VLC domain. Utilizing the intra prediction mode and coded block pattern in VLC domain, the proposed algorithm chooses tho...A semi-fragile content authentication algorithm is proposed for low bit-rate H.264/AVC video in VLC domain. Utilizing the intra prediction mode and coded block pattern in VLC domain, the proposed algorithm chooses those macro-blocks from which the signature is extracted and constructs content signature at macro-block level according to the relationship among the energies of quantized low-frequency coefficients of sub-macroblocks. The signature is embedded by modifying the trailing coefficients. The experimental results show that the proposed algorithm performs well in visual quality impact and keep the bit-rate basically unchanged. In addition, the algorithm can embed signatures into I, P, B slices simultaneously and remarkably enhances the watermark capacity. By verifying the extracted signature, the algorithm can detect and locate video tampering efficiently.展开更多
基金This paper is sponsored by the National Natural Science Foundation of China （No. 60802057, 61071153）, National 863 Plan of China （ 2009AA01Z407 ）, Shanghai Rising-Star Program （10QA1403700）, and Shanghai Educational Development Foundation.
文摘A semi-fragile content authentication algorithm is proposed for low bit-rate H.264/AVC video in VLC domain. Utilizing the intra prediction mode and coded block pattern in VLC domain, the proposed algorithm chooses those macro-blocks from which the signature is extracted and constructs content signature at macro-block level according to the relationship among the energies of quantized low-frequency coefficients of sub-macroblocks. The signature is embedded by modifying the trailing coefficients. The experimental results show that the proposed algorithm performs well in visual quality impact and keep the bit-rate basically unchanged. In addition, the algorithm can embed signatures into I, P, B slices simultaneously and remarkably enhances the watermark capacity. By verifying the extracted signature, the algorithm can detect and locate video tampering efficiently.
