【摘要】 为了更好地保护音频作品的所有权,提出了一种基于LWT-SVD混合域的鲁棒性音频水印算法。借助混沌和置乱为图像水印加密,保证水印安全性;利用提升小波变换(LWT)和奇异值分解(SVD)的特性,在音频LWTSVD的奇异值中嵌入式水印信息,保证水印的鲁棒性;此外,借助汉明码纠错进一步提升水印的鲁棒性。仿真实验结果表明,本算法的水印安全性高、透明性好,对一般音频信号处理鲁棒性强。更多还原

【Abstract】 In order to better protect the ownership of audio works,a robust digital audio watermarking algorithm in the hybrid wavelet domain is presented,combined with the important features of lifting wavelet transform and singular value decomposition.First of all,the algorithm encrypts the binary watermark image twice by means of image scrambling and chaotic sequence to ensure the security of watermark.Secondly,the audio is divided into some frames according to the rows of the watermark,and each frame is carried on multi-level lifting wavelet decomposition.Low frequency components of the wavelet domain are selected for singular value decomposition.Then the encrypting watermark information is embedded in the singular value of the audio to ensure the watermark robustness.In addition,the algorithm corrects the errors by hamming encoding,further enhancing the robustness of the watermark.The simulation results show that watermarking security and transparency of the algorithm are very good,for it has good robustness in the general audio processing of Gaussian noise,low-pass filtering,resampling,requantization,and lossy compression.The algorithm doesn′t need the original carrier audio and the binary watermark image in extraction process,really realizing blind extraction ofwatermark.Therefore,this algorithm has high value in the practical applications.更多还原