基于DWT和均值量化的音频水印算法

Audio watermarking scheme based on DWT and mean quantization

提出基于小波包的离散小波变换和均值量化的音频水印算法。该算法的主要特点是:1)将音频信号根据嵌入的水印图像大小自适应地分段。利用Arnold变换对水印图像进行预处理,消除水印图像的相关性,提高图像的抗攻击性。2)采用小波包的离散小波变换,增加嵌入和提取效率,提高水印的透明性。3)水印信息嵌入到均值中,有较好的不可感知性。仿真实验表明:嵌入水印后音频的不可感知性较高,能够有效地抗击有损压缩、低通滤波和剪切等攻击。

This paper presented a blind digital audio watermarking scheme based on wavelet packet and mean quantization.The features of proposed scheme are as follows:1) Audio signals were decomposed into many segments according to the size of the image watermarking.The image watermarking was preprocessed using Arnold transformation to have a strong robustness to some attack.2) Discrete wavelet transform(DWT) was used to improve the transparency of digital watermarking.3) The image watermarking was embedded into the m...