一种基于ICA的音频数字水印方法  

A Method of Digital Audio Watermarking Based on Independent Component Analysis

作　　者：李亚琴[1] 孙星明[2] 

机构地区：[1]湖南大学软件学院,湖南长沙410082 [2]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《计算机工程与科学》2007年第7期41-44,共4页Computer Engineering & Science

基　　金：国家自然科学基金资助项目(60373062);教育部科研重点项目(03092);湖南省杰出中青年基金资助项目(02JJYB012)

摘　　要：本文提出了一种基于独立分量分析ICA的音频数字水印方法。嵌入水印时,对原始音频文件进行一级小波分解,在分解得到的逼近分量中利用随机混合方法嵌入水印后进行小波逆变换形成嵌入水印的音频文件。提取水印时,对音频文件进行一级小波分解,用基于最大峰值的ICA方法进行水印提取。仿真实验结果表明,这种方法的强鲁棒性和可行性很强。A method of digital audio watermarking based on independent component analysis is proposed in this paper. When embedding watermarks, the original audio is decomposed by a single-level discrete 1-Dwavelet transform, and watermarks are embedded in the obtained approximation coefficient vector by a method of random mixing,and an audio file is reconstructed by a reverse wavelet transform. When watermarks are extracted, the audio file is decomposed by a single-level discrete 1-Dwavelet transform, and watermarks are extracted by ICA based on Kurtosis maximization. Emulated experimental results show that this method is robust and feasible.

关 键 词：音频数字水印 独立分量分析 离散小波变换 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]