Copyright infringement is a serious problem in the music industry, especially for Internet-based platforms. The existing audio watermarking methods for copyright protection have limited robustness against the desynchronization attacks. Besides, most of the methods are not designed for stereo signals, which are widely used by online-based music and video platforms. In this paper, we present a desynchronization-attack-resilient watermarking method for stereo signals. In the watermark embedding process, both the left channel and the right channel of the host signal are firstly segmented and transformed into the frequency domain using the discrete cosine transform (DCT). Secondly, the linear correlation between channels (LCBC) in the DCT domain is calculated using the Pearson correlation coefficient (PCC). Then watermark bits are embedded by modifying LCBC for each segment. Since the desynchronization attacks as well as other common attacks affect both the left channel and the right channel in a similar manner, LCBC is robust against both desynchronization attacks and other common attacks. Theoretical analysis and experimental results demonstrate the effectiveness of the proposed method.

版权侵权是音乐行业的一个严重问题，尤其是对于基于互联网的平台。现有的用于版权保护的音频水印方法对去同步攻击的鲁棒性有限。此外，大多数方法不是为立体声信号设计的，立体声信号被在线音乐和视频平台广泛使用。在本文中，我们提出了一种用于立体声信号的去同步攻击弹性水印方法。在水印嵌入过程中，首先对宿主信号的左声道和右声道进行分割，并使用离散余弦变换（DCT）将其变换到频域中。其次，使用Pearson相关系数（PCC）计算DCT域中信道间的线性相关（LCBC）。然后通过修改每个段的 LCBC 来嵌入水印位。由于去同步攻击以及其他常见攻击以类似的方式影响左声道和右声道，因此 LCBC 对去同步攻击和其他常见攻击都具有鲁棒性。理论分析和实验结果证明了该方法的有效性。