摘要
为了实现音频数字媒体的认证和版权的合法保护,提出了一种新的基于离散小波变换(DWT)和奇异值分解(SVD)的音频零水印算法。将音频分段,对分段后的音频信号进行二级离散小波变换,对变换后的近似分量进行奇异值分解,利用对角矩阵产生x(i)序列,构造二值水印密钥,密钥和水印异或得到检测密钥。仿真结果表明,该算法对重采样、低通滤波、重量化、降噪、MP3压缩、回声等常见音频信号处理攻击具有很强的鲁棒性和透明性。
To realize the authentication of audio digital media and legal protection of audio digital media copyright, a novel audio zero watermark algorithm based on DWT-SVD is proposed. In the algorithm, the audio is divided into sections. Each audio sec tion is decomposed on two-level discrete wavelet transform. Then the approximate coefficients are decomposed on singular value decomposition transform. A sequence number x（i） is gotten according to the characteristic of the diagonal matrix and binary wa termark key is constructed. Experimental results demonstrate that the proposed watermarking algorithm has strong robustness against popular attacks such as re-sampling, low-pass filter, Gaussian white noise, re-quantization, noise reduction, MP3 com pression and echo.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第1期42-46,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61163066
60902074)
全国教育科学规划课题基金项目(FFB108172)
关键词
数字水印
音频水印
零水印
离散小波变换
奇异值分解
digital watermark
audio watermark
zero watermark
discrete wavelet transform
singular value decomposition transform
