【摘要】 提出一种新的、基于支持向量回归(SVR)的音频水印提取方法。在该方法中,利用子采样技术将原始音频信号划分为四个子音频信号。这些子音频对应的DCT系数间具有高的相关性,将这种相关性视为一种回归问题。在水印提取过程中,利用支持向量回归来学习这种相关性,并使用训练好的SVR完成水印的提取。实验结果表明该方法对比其它几种方法有良好的性能和更好的鲁棒性。更多还原

【Abstract】 A new watermark extraction method for audio signal based on support vector regression(SVR) is presented in this paper.Four sub-audio signals are obtained from original audio signal by using subsampling technique.The close relativity between the corresponding coefficients in the four sub-audio signals is viewed as a regression problem.In watermark extraction procedure,the relativity can be learned by the training procedure of support vector regression,and then the extraction of watermark is completed by trained SVR.Experimental results show that the proposed method has good performance and the robustness of the proposed method is superior to other several methods.更多还原