基于随机共振的鲁棒性音频数字水印算法

Robust audio watermark algorithm based on stochastic resonance

为了提高应用于版权保护的音频数字水印算法的鲁棒性,提出了一种基于音频混合域随机共振的数字水印算法。在嵌入端,将水印序列嵌入在音频载体的混合域系数中。在检测端,待检测音频的混合域系数在预处理后,作为随机共振系统的输入,通过调节随机共振信号处理器的参数和本文提出的去刺参数以产生随机共振从而实现水印的检测。仿真结果表明,该算法能够在保证水印信息不可感知的前提下抵抗StirmarkForAudio软件的多种攻击,相比常见算法显示出更强的鲁棒性。

To improve robust ofthe audio digital watermarking algorithm used in copyright protection, a new audio watermark algorithm based on stochastic resonance in the DWT and DCT domain is proposed. In embed-process, we embed the watermark sequence in the DWT and DCT domain coefficients of audio carriers. In decode-process, after the preprocessing of the audio to be decoded, the DWT and DCT domain coefficients are used as the input of the stochastic resonance system, by adjusting the parameter of the stochastic resonance processor and the deburring parameters which presented, the detection of watermark is achieved. Simulation results show that it can resist most attacks in StirmarkForAudio with unperceived watermark, which shows the stronger robustness of the proposed approach than common algorithms.