基于混沌的离散小波变换鲁棒音频盲水印  

Robust Audio Blind Watermarking of Discrete Wavelet Transform Based on Chaos

作　　者：黄雄华[1] 蒋伟贞[2] 

机构地区：[1]桂林电子科技大学教学实践部,广西桂林541004 [2]暨南大学信息科学技术学院,广东广州510084

出　　处：《成都大学学报（自然科学版）》2009年第4期318-321,共4页Journal of Chengdu University（Natural Science Edition）

摘　　要：提出了基于混沌的离散小波变换(DWT)的数字音频盲检测算法.算法首先对原始音频数据进行分段处理,然后对每个音频数据段进行4级DWT变换,最后利用4级细节小波系数统计均值趋于零的特性,通过修改4级细节小波系数统计均值,每一段中嵌入一个水印比特,算法提取水印时不需要原始音频信号.为了便于观察,水印选用了二值可视图像,并利用混沌技术进行了加密处理,加强其安全性.仿真实验证明了算法对滤波、有损压缩、重采样等常见信号处理方法具有鲁棒性,与同类算法相比,具有更好的不可感知性.An algorithm for digital audio blind watermarking of discrete wavelet transform （DWT） was proposed based on chaos. Firstly, the original audio signal was divided into several audio segments; then, fourlevel DWT was performed on every audio segment; finally, based on the property that the statistical mean value of 4th level wavelet coefficients is next to zero, a bit watermarking was embedded into a segment, by modifying the statistical mean value of the 4th level detail wavelet coefficients. The watermarking could be extracted without the original audio signal. In order to observe conveniently, two-value image watermarking was selected and the chaos technology was used to encrypt the image for the sake of security. The simulated experimental results prove that the algorithm is robust to methods for common signal processing, such as wave filtering, lossy compression, resample and so on. Compared with similar algorithms, this algorithm has better imperceptibility.

关 键 词：音频水印 统计均值 混沌 盲检测 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]