一种DCT域自适应音频水印算法  

Adaptive audio watermarking algorithm in DCT domain

作　　者：黄雄华[1,2] 王宏霞[1] 

机构地区：[1]西南交通大学信息科学与技术学院,成都610031 [2]桂林电子科技大学教学实践部,广西桂林541004

出　　处：《计算机应用研究》2009年第8期2989-2991,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60702025);高等学校博士学科点专项科研基金资助项目(20070613024);四川省青年科技基金资助项目(07ZQ026-004)

摘　　要：数字水印技术中水印强度是影响水印鲁棒性的重要参数。当前大多数的水印算法均利用实验来确定水印强度,但实验具有很大的随机性,得到合适的水印强度需要大量的实验。提出了一种DCT(discrete cosinetransform)域基于信噪比(signal noise ratio,SNR)水印强度自适应的音频水印。首先将载体音频分帧,然后根据对嵌入水印后音频品质要求的SNR计算各帧的水印强度,选择可嵌入强度大的帧,将一个水印比特信息嵌入一帧的DCT的直流系数,水印强度将自适应于各帧。实验结果证明了该算法具有较好的鲁棒性。本算法可以由SNR自适应调控嵌入强度,从而取得水印鲁棒性和不可感知较好折中的结论。The watermarking strength in digital watermarking technology is an important parameter affecting the robustness. Most of watermarking algorithms nowadays make use of experiment to obtain certain the watermarking strength. The results of experiment will be random. So, a lot of experiments have to be done in order to obtain appropriate strength factor. This paper proposed an adaptive audio watermarking strength based on SNR in DCT（ discrete cosine transform）domain. Firstly, divided carrier audio into some frames, and then calculated watermarking strength according to SNR which was required by the quality of watermarked audio. Chose the frames that allowed larger strength and embedded a watermarking bit into a DCT direct current coefficient chosen frame. Thus, watermarking strength would be adaptive to the frames. Experimental results demonstrate the algorithm has good robustness. Obtains the conclusion that the algorithm can control watermarking strong by SNR and make good tradeoff of watermarking＇ s roubustness and imperceptibility.

关 键 词：音频水印 离散余弦变换 信噪比 自适应 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]