基于能量量化的小波域数字音频水印算法  

A DWT Domain Digital Audio Watermarking Algorithm Based on Energy Quantization

作　　者：王庆席[1] 郑晓势[2] 刘广起[2] 赵彦玲[2] 李娜[2] 

机构地区：[1]山东轻工业学院信息科学与技术学院,济南250100 [2]山东省计算中心,济南250014

出　　处：《信息技术与信息化》2006年第3期75-77,共3页Information Technology and Informatization

摘　　要：利用混沌动力学系统中logistic映射先对水印图像进行置乱预处理以增强其鲁棒性,在小波域中将原始音频信号按比例分段,求得每段的能量然后将其按步长delta以及要嵌入的水印的值进行量化,然后用量化后的能量与量化前的能量之比乘以该段每一个值即完成一位水印的嵌入。反复执行以上步骤,完成全部水印的嵌入。若低频部分分组数大于水印序列个数则重复嵌入。采用StirMarkforAudio攻击后的实验结果表明该方案对StirMarkforAudio的部分攻击具有一定的鲁棒性。First we use the logistic-map sequences to encrypt the watermark, then we separate the original audio signal in DWT domain into some pieces by the rate you wanted. In every piece we calculate its energy and quantize the energy by the delta you set and the value of the corresponding bit of watermark., and we can get the relative value of them. We add one bit of the watermark in the DWT domain by muliting the ratio in one piece. We repeat the above steps until all the bits are embeded. We can re-embed the watermark if the total number of the separated groups is bigger than that of watermark. Experimental results show that it is robust against some StirMark attacks.

关 键 词：能量量化 置乱加密 离散小波变换 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]