摘要
提出了一种基于奇异值分解和离散小波变换的音频盲水印方法,将原始音频数据进行小波变换,提取低频分量分块奇异值分解后,利用置乱加密处理的水印图像,量化选取的奇异值,实现水印的嵌入过程.利用图像置乱和奇异值分解方法,提高了水印的不可感知性.另外在水印的检测与提取过程中不再需要原始载体数据,更加利于实际中的应用.通过MATLAB仿真实验表明,水印隐藏效果好,而且对多种攻击具有较好的鲁棒性.
A blind watermarking scheme is proposed based on SVD-DWT （SVD Singular Value Decomposition, DWT Discrete Wavelet Transform）. The audio signal is decomposed using appropriate wavelet basis. Low frequency coefficients are selected to divide into section and SVD of the divided signals is made. The singular values are chosen and embedded into the watermark image by quantization method. The using of singular value decomposition and discrete wavelet transform methods improves the imperceptible and the security of the watermark. The watermark can be extracted without the original digital audio signal. Experimental results show that method in this paper is robust to many attacks, such as resampling,low pass filtering, noise addition and so on.
出处
《湖南师范大学自然科学学报》
CAS
北大核心
2009年第1期47-50,55,共5页
Journal of Natural Science of Hunan Normal University
基金
湖南省教育厅科技研究基金资助项目(07C577)
关键词
奇异值分解
音频水印
离散小波变换
ARNOLD变换
singular value decomposition
audio watermarking
discrete wavelet transform
Arnold transform
