摘要
空间域随机置乱和锐化滤波是基本的图像处理方法。在某些数字图像水印系统中,水印嵌入前对原始图像做随机置乱预处理,水印提取前对水印图像做锐化滤波预处理,可以有效地改善视觉效果,提升水印检测的准确率。本文在Barni半盲水印算法中验证了该方法的有效性。此外,本文设计了一个新的DCT域水印算法,并对该方法做了进一步验证。算法中,将水印嵌入到图像相邻块的DCT系数差值上。实验证明,该算法在保持良好视觉效果的前提下对常见图像处理具有较强的鲁棒性。
Random permutation and sharpening are basic methods of image processing. In some digital image watermarking systems, if the original image is pre-processed with random permutation before embedding and the watermarked image is pre-processed with sharpening before detection, the performance of watermarking can be improved, which means the images can look better, or the veracity rate of detection is higher. In this article, this scheme is validated in Bami＇s semi-blind watermarking system. In addition,a new watermarking scheme is designed in which both permutation and sharpening are used to insure the performance. In this scheme, watermarking is embedded into the difference of neighboring blocks＇ DCT coefficients. The experimental results show that this scheme is robust to various image processing.
出处
《计算机工程与科学》
CSCD
2008年第9期47-50,共4页
Computer Engineering & Science
基金
国家自然科学基金资助项目(60502023)
湖北省自然科学基金资助项目(CGZ0223)
关键词
数字水印
盲检测
扩频通讯
随机置乱
锐化滤波
digital watermarking
blind detection
spread spectrum communication
random permutation
sharpening
作者简介
徐端全（1969-），男，博士生，讲师，研究方向为图像／视频编码、数字水印技术等；通讯地址：430074湖北省武汉市珞喻路1037号华中科技大学电信系宽带多媒体中心；Tel：（027）87792463
朱光喜，教授，博士生导师，研究方向为宽带无线与多媒体通信。
