一种新的数字音频多水印算法  

A New Digital Audio Multi-Watermarking Algorithm

作　　者：徐挺挺[1] 杨震[1] 

机构地区：[1]南京邮电大学信号处理与传输研究院,南京210003

出　　处：《信号处理》2009年第3期444-449,共6页Journal of Signal Processing

基　　金：国家863重点项目(2006AA010102);江苏省自然科学基金项目(BK2004150)资助课题

摘　　要：本文以同时实现数字媒体的版权保护和内容认征为目标,提出一种在同一个音频文件中嵌入三个不同水印的数字音频多水印算法。鲁棒水印经置乱预处理后,通过SCS方法(Scalar Costa Scheme)嵌入音频信号的低频小波系数中,具有良好的鲁棒性。它不仅对加噪、低通滤波等常见信号处理稳健,还具有抵抗恶意替换攻击的能力,实现了音频文件的版权保护。时域SCS算法嵌入的半易损水印对一般信号操作稳健而对恶意篡改敏感并可以对篡改区域进行定位。最后通过改进的融合信息隐藏算法在时域嵌入全易损水印,它对加之于音频信号上的任何微小变化都非常敏感。可对原始文件的完整性进行认证。实验结果表明,整个系统透明性良好。三种水印提取时均不需提供原始音频信号,具有一定的实用价值。With the purpose of protecting the copyright of digital media and verifying its originality, the paper proposes a new digital audio multi-watermarking scheme which embedded three different watermarks in the same audio file. After the pretreatment of scrambling, the robust watermark is embedded in the DWT low frequency coefficients by SCS algorithm. It is robust to signal attacks such as noise adding and low-pass filtering as well as malicious substitution attack thus can be applied to copyright protection. The semifragile watermark embedded in time domain by SCS algorithm is robust to common signal operations but sensitive to malicious substitution attack and can locate the substitution area while the fragile watermark embedded by improved amalgamation algorithm is sensitive to any small changes of the audio signal thus can verify the integrity of the original audio file. Experimental result shows that the transparency of the system is satisfying. All of the three watermarks are blind-extracting, which is useful in practice.

关 键 词：数字音频多水印 盲提取 鲁棒水印 全易损水印 半易损水印 

分 类 号：TN912.3[电子电信—通信与信息系统]