基于混沌的复倒谱域双重音频水印算法

为了增强水印的安全性,采用混沌理论对水印信息进行加密,在无法获知加密密钥的情况下,水印信息不能被完全获取。基于时域能量特点自适应选取嵌入帧,针对音频信号倒谱变换后倒谱系数的特征,利用统计均值调制的方法进行嵌入,提高了水印的不可感知性。并引入了多重水印的思想,在抵抗多重水印攻击的情况下,也给数字音频版权保护带来帮助。实验证明,该算法在随机剪切、抖动攻击等同步攻击下具有良好的鲁棒性。

Doubled Audio Watermarking Algorithm based on Chaos in Complex Cepstrum Domain

To enhance the security of watermark,chaos theory is used to encrypt the watermark information,and if the key isn’t known,the watermak information could not be attained completely.The adaptive selection of embedded frame based on characteristics of the energy in time-domain is done,the watermark is embedded by SMM(Statistical Mean Modulation) according to its cepstral features after the audio signal is transformed to complex cepstrum domain.Non-perceptibility of the watermark is improved with this approach.In addition,the idea of multiple watermarking is introduced,which could contribute to the copyright protection of digital audio in resistance of multiple watermarking attacks.Experimental results show that this algorithm has a pretty robustness under various synchronous attacks such as random cutting,vibrating attack and so on.