基于小波变换和倒谱变换的音频水印技术  

An Audio Watermarking Algorithm Based on Wavelet and Cepstrum Transform

作　　者：禹召阳[1,2] 张晓明[2] 

机构地区：[1]北京化工大学,北京100029 [2]北京石油化工学院,北京102617

出　　处：《电脑知识与技术》2009年第6期4271-4273,共3页Computer Knowledge and Technology

摘　　要：提出了一种基于小波变换和倒谱变换的数字音频水印算法。该算法将音频信号进行小波分解，选取低频系数。对低频系数进行倒谱变换，并根据统计均值思想，把水印嵌入到倒谱域中，实现了在音频信号中嵌入二值图像。实验结果表明该算法具有比较好的鲁棒性，能抵抗常见信号的攻击，如低通滤波、重采样、重量化、有损压缩、音量变化、添加白噪声等。An audio watermarking algorithm is proposed based on wavelet and cepstrum transform. It decomposes the audio signal in wavlete domain, by extracting the wavelet approximate components and selecting the low-frequent coefficient. Then the low-frequent coefficient is transformed in cepstmm domain. According to the statistical mean manipulation idea, an image watermark is embedded into the cepstrum domain of the audio signal. Experimental results show that the algorithm has better robustness on resisting Common signal processing such as low-pass faltering, resampling, re-quantization, lossy compression, the volume change and adding white noise.

关 键 词：音频水印 小波变换 倒谱变换 统计均值 鲁棒性 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]