基于混沌加密的盲音频水印算法

将混沌理论引入到音频水印系统的研究中，设计了一种包含版权信息和密钥的二值图像作为水印，利用混沌序列将其加密后嵌入到原始音频信号的DWT域中。该水印方案包含的版权信息量大，且具有双重的安全性，很难提取出水印信号；即使提取出水印信号，如果不知道密钥，也很难恢复出原始水印图像。本算法采用了量化方法嵌入水印，提取水印时不需要原始的音频信号，实现了水印的盲提取。对一般的信号处理进行了试验，水印具有很强的鲁棒性。

Audio Watermarking Algorithm in Wavelet Domain Based on Chaotic Encryption

The technology of embedding image into the audio signal was researched. A chaotic encrypted binary image watermarking could be embedded in the original signal wavelet domain in the algorithm. The watermark contains much copyright information and has dual security. It is difficult to extract the watermark signal,and even if the watermark was extracted,if the key isn't known,it is hard to recover the original watermark. The quantitative method is used in watermarking embedding,and the watermark signal can be extracted without access to the original audio signal. This blind watermark algorithm has practical value. The experimental results show that the watermarks are imperceptible and the algorithm is robust to many operations of digital audio signal,such as resampling,cropping, low pass filter,MP3 compression and so on.