基于混沌扩频和能量系数差的音频盲水印算法

提出一种基于混沌扩频和能量系数差的小波域数字音频盲水印算法。采用混沌扩频序列对原始水印进行加密，小波算法对原始音频信号进行分解，引入心理声学模型选取适当的阈值，根据高频系数分量和低频系数分量的能量差值与阈值的大小关系，实现对水印信息的嵌入、提取和盲检测。在提取过程中结合线性伸缩恢复的方法消除时间轴上线性伸缩带来的影响。实验结果表明，该算法对多种音频文件的操作和攻击均具有良好的鲁棒性。

Blind Audio Watermark Algorithm Based on Chaotic Spread Spectrum and Difference of Energy Coefficients

This paper proposes a blind audio watermark algorithm based on chaotic spread spectrum and difference of energy coefficients during the research of audio watermark technique. The watermark data is encrypted by Chebyshev chaotic spread spectrum based on secret key. Audio signal is decomposed by wavelet transform and the energy differences of the approximation coefficients and detail coefficients are collected and classified into two classes according to the relationship between the energy differences and the thresholds based on psychoacoustics model. In this way, the watermarks are embedded, extracted and achieved not only the blind signal detection, but also greater degree of increase watermark robustness. The restoration of the linear scaling is used to resist the influence of the linear scaling on time axis. Experimental results show that this algorithm is robust to many operations and attacks on digital audio signal.