抗重录音的溯源追踪音频水印算法

AUDIO WATERMARKING ALGORITHM FOR TRACING THE RE-RECORDED AUDIO SOURCE

录音设备的普及,使重录音成为一种去除音频信号中嵌入水印的普遍攻击方式。为抵抗重录音攻击提出一种新的用于溯源追踪的鲁棒音频水印算法,给出系数对数均值(Logarithmic Mean of Coefficient,LMC)的定义,分析系数对数均值的性质。提出基于DCT的中频域水印嵌入方法。同步信息由帧号生成,和由版权生成的水印信息一起嵌入到载体信号中。仿真实验结果表明,提出的算法具有较好的不可听性、鲁棒性,同时能够有效抵抗一定程度的重录音攻击。

With the popularity of recording equipment,re-recording becomes a popular attack method to remove watermarks embedded in audio signals.In order to resist re-recording attacks,a new robust audio watermarking algorithm for source tracking is proposed.The definition of logarithmic mean of coefficient(LMC)was given,and the properties of the logarithm mean of coefficients were analyzed.A DCT-based middle frequency domain watermark embedding method was proposed.The synchronization information was generated after the frame number was binarized,and the frame number and the watermark were embedded in the carrier signal audio frame.The simulation results show that the proposed algorithm has good imperceptibility and robustness,and can effectively resist a certain degree of re-recording attacks.