计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (32): 55-58.

• 学术探讨 • 上一篇    下一篇

小波域基于MFA_ICA检测的数字图像水印算法

张政保1,杨素敏1,籍风荣2,丁国良1   

  1. 1.军械工程学院 计算机工程系,石家庄 050003
    2.军械工程学院 导弹工程系,石家庄 050003
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-11-11 发布日期:2007-11-11
  • 通讯作者: 张政保

Algorithm for image watermark detection based on MFA_ICA in wavelet domain

ZHANG Zheng-bao1,YANG Su-min1,JI Feng-rong2,Ding Guo-liang1   

  1. 1.Department of Computer Engineering,Ordnance Engineering College,Shijiazhuang 050003,China
    2.Department of Missile Engineering,Ordnance Engineering College,Shijiazhuang 050003,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-11-11 Published:2007-11-11
  • Contact: ZHANG Zheng-bao

摘要: 基于MFA_ICA(Independent Component Analysis based on Mean Field Approximation)理论提出了一种数字图像水印算法。该算法利用MFA_ICA进行水印和原始图像的估计,并用估计出的独立分量检测水印是否存在。与传统线性相关检测器相比,该算法消除了原始载体的干扰,检测性能明显提高。重要的是,不仅通过仿真实验证明了该算法较强抵抗JPEG压缩、剪切和白噪声等攻击的能力,而且从理论上也给出了严格的证明。

Abstract: Due to the ability to detect watermark without the original image,the process of blind watermarking is similar to blind signal separation,so in the paper an algorithm for blind image watermark detection based on MFA_ICA is proposed.At first,the proposed blind watermark detector employs MF_ICA for the hidden independent components(i.e.the watermark and the host signal) from watermarked image,then the estimated components are used to detect the presence of the embedded watermark.The proposed algorithm is capable of not only reducing the interference of the host signal,but also canceling the influence of the decoding error probability.The most important is that the performance of the proposed algorithm is not only verified by kinds of attacking experiments,such as the JPEG compression,cropping and adding white noise,but also proved by theoretical deduction.Analysis results are presented showing that the proposed detector performs significantly better than the linear correlation detector.