基于SVR的混合域鲁棒音频水印算法  

Robust Audio Watermarking Algorithm in Hybrid Domain Based on SVR

作　　者：齐燕博 杜庆治[1] 龙华[1] 邵玉斌[1] QI Yanbo;DU Qingzhi;LONG Hua;SHAO Yubin(College of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《电视技术》2022年第2期14-20,共7页Video Engineering

摘　　要：对于常见的基于离散小波变换-奇异值分解的水印算法应对常见的攻击鲁棒性较差的情况,提出一种基于奇异值比(Singular Value Ratio,SVR)的混合域音频水印算法。先将音频分帧,利用音频特性选择适合水印嵌入的帧,然后对水印嵌入帧进行离散小波变换和离散余弦变换,将离散余弦变换系数分为4段,选取中频系数进行奇异值分解,计算奇异值比。根据音频的质量最优修改奇异值,嵌入水印。同时利用Arnold变换和Logistic混沌序列提升其安全性,利用汉明码增强其纠错能力。实验表明所提算法安全性高,对添加高斯噪声、低通滤波、重采样、重量化和压缩具有良好的鲁棒性。For the common watermarking algorithms based on discrete wavelet transform singular value decomposition, in order to deal with the poor robustness of common attacks, this paper proposes a hybrid domain audio watermarking algorithm based on Singular Value Ratio(SVR), which uses the audio characteristics to select the frame suitable for watermark embedding to realize the dynamic embedding of watermark. Firstly, the audio is divided into frames, then discrete wavelet transform and discrete cosine transform are carried out,and then the discrete cosine transform coefficients are divided into two segments for singular value decomposition, and then SVR is calculated. The singular value is modified according to the optimal audio quality. At the same time, Arnold transform and Logistic chaotic sequence are used to improve its security. Hamming code is used to enhance its error correction ability. Experiments show that this algorithm has high security and good robustness to Gaussian noise, low-pass filtering, resampling, weighting and compression.

关 键 词：数字音频水印 奇异值比 离散小波变换 离散余弦变换 ARNOLD变换 

分 类 号：TP311.1[自动化与计算机技术—计算机软件与理论]