Description: 提出了基于Contourlet变换的数字图像水印算法。与小波变换不同的是，Contourlet变换采用类似于线段
（contour segment）的基得到一种多分辨、局部化、方向性的图像表示。水印信号通过基于内容的乘性方案加载
到 Contourlet 变换系数。在采用零均值广义高斯分布拟合 Contourlet 变换系数的基础上，提出采用极大似然估计
实现水印的盲检测。依据 Neyman-Pearson 准则，在给定虚警率的情况下对判决准则进行了优化。实验结果表明
在保证水印隐蔽性的前提下，水印对常见的信号处理手段以及几何变换具有很好的稳健性-Contourlet transform is proposed based on digital image watermarking algorithm. The difference is that with the wavelet transform, Contourlet transformation similar to that of line (contour segment) of the base by a multi-resolution, localization, directional image representation. Watermark-based content of the program is loaded into Contourlet by transform coefficients. The introduction of zero-mean generalized Gaussian distribution fitting Contourlet transform coefficient based on maximum likelihood estimation proposed by the blind watermark detection. Based on Neyman-Pearson criterion, the false alarm rate in a given case to the sentencing guidelines were optimized. The results show that the watermark hidden in the guarantee under the premise of the watermark signal processing means of the common geometric transformations and is robust Platform: |
Size: 134144 |
Author:betty |
Hits:
Search - difference of gaussian in image processing