基于最优小波包变换和离散余弦变换的灰度图像水印算法  

The algorithm of gray image watermark based on best wavelet packet and discrete cosine transform

作　　者：许慧[1] 谭阳红[1] 孙磊[1] 龙伯华[1] 

机构地区：[1]湖南大学电气与信息工程学院,湖南长沙410082

出　　处：《电路与系统学报》2009年第1期97-101,共5页Journal of Circuits and Systems

基　　金：国家自然科学基金(50277010);湖南省自然科学基金(03JJY1010;04FJ2003;05GK2005);教育部博士基金(20020532016);高等学校新世纪优秀人才支持计划(NCET-04-0767)

摘　　要：本文提出了一种基于最优小波包变换和离散余弦变换的灰度图像水印算法,利用Arnold变换和Baker映射对水印图像进行置乱,并将置乱结果进行分块,对宿主图像进行最优小波包分解,然后修改低频子带系数来嵌入水印信息。嵌入前,利用离散余弦变换去除低频子带相邻系数的相关性后嵌入经过同样处理的置乱水印子块。实验表明该算法在抵抗噪声、滤波、旋转剪切等攻击下有比较好的鲁棒性。Based on best wavelet packet transform （BWPT） and discrete cosine transform （DCT）, a novel algorithm of gray image watermark is proposed. Watermark image is scrambled through Arnold transform and Baker mapping, and thcn is divided into four blocks. The host image is decomposed based on BWPT. The process is completed by modifying low frequency band coefficients of host image. Before watermark cmbedcd, DCT is used to reduce the correlation of the low frequency band adjacent coefficients of host image, then watermark-blocks operated by the same process arc embedded. The results show the proposed scheme is robust against noise; filter, rotation etc.

关 键 词：小波包变换 离散余弦变换 数字水印 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]