摘要
基于盲信号分离技术,提出了一种新的数字水印图像提取方法.该方法在一种较典型的基于小波变换的数字水印嵌入算法基础上,应用盲分离技术有效地检测和提取出了含在水印图像中的水印.经大量试验证明,所提出的方法对常用的图像处理方法有较强的抗攻击性能.
Blind source separation can separate or extract the source signals from mixing signals, without knowing the characteristics of the transmission channels and source signals. A image watermark detection and extraction technique based on blind source separation was proposed. After introducing a general image watermarking embedding method in wavelet domain, the paper applies blind separation algorithms using cumulants for watermark detection and extraction in every wavelet domain. The experimental results show the proposed algorithm is robust.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2004年第2期229-232,共4页
Journal of Shanghai Jiaotong University
关键词
盲分离
数字图像处理
数字图像水印
小波变换
检测
提取
Blind source separation
Communication channels (information theory)
Computer simulation
Feature extraction
Image processing
Wavelet transforms
