摘要
结合视频压缩标准、脆弱数字水印及视频通信技术,文中提出一种有效识别正版视频的TDIA算法.它包括嵌入算法和识别算法两部分.为消除通信干扰对水印的影响,将迭代方法引入识别算法.基于盗版操作与通信干扰的本质不同,研究讨论了迭代方法的性能.最后,基于迭代结果,借助Chernoff Bound理论,分析了TDIA算法的识别误差.对于码流BER(Bit Error Rate)10-3、视频I帧总数n360的Mpeg2编码视频,该算法识别误差小于10-18.
An efficient genuine-video identification algorithm named as TDIA is proposed, which based on integration of video compression standards, fragile watermarking and video communication technology. It includes two parts, embedding algorithm and identification algorithm. Iterative method is introduced on the identification algorithms to eliminate the communications interference for the impact of watermarking. The performance of the iterative method is investigated according to the difference of nature between piracy operation and communications interference. Finally, the identification errors of TDIA algorithm are analyzed based on the results of iteration and by dint of the theory of Chernoff Bound. The total identification error of the algorithm is less than 10^-18 when BEN （Bit Error Rate） of encoded video stream is not more than 10^-3 and the total number of I-frames of Mpeg2 encoding video is more than 360.
出处
《计算机学报》
EI
CSCD
北大核心
2010年第1期175-183,共9页
Chinese Journal of Computers
