鲁棒音频数字水印图频域算法研究

A Robust Audio Digital Watermarking Algorithm in Graph Frequency Domain

为了解决音频水印的鲁棒性问题,本文提出一种基于图傅里叶变换和奇异值分解的音频水印算法。该算法借助Arnold对图像水印进行置乱加密,利用图傅里叶变换和奇异值分解的特性,将加密后的图像水印嵌入到音频载体上。仿真实验结果表明,音频载体在经过噪声、压缩、重采样、滤波、压缩等多种攻击后能提取出较完整的图像水印,证明该算法具有良好的不可感知性、保密性和鲁棒性。

In order to solve the robustness problem of audio watermarking,this paper proposes a robust audio watermarking algorithm based on Graph Fourier transform and singular value decomposition.With the help of Arnold,the image watermark is scrambled and encrypted.The encrypted image watermarking is embedded into the audio carrier by utilizing the characteristics of Fourier transform and singular value decomposition.The simulation results show that the watermark can extract a complete image watermark after noise,compression,resampling,filtering,compression and other attacks.The algorithm has good invisibility,confidentiality and robustness.