摘要
文中提出了一种基于小波变换和神经网络的盲数字水印算法。算法嵌入的水印信息是一幅二值图像,每位水印信息被连续多次嵌入宿主图像小波分解的低频子带。利用提取含水印图像小波分解的低频子图的特征点坐标集估计水印图像遭受的几何失真,根据估计的几何失真参数对受到几何攻击的水印图像进行校正,最后利用神经网络很好的非线性映射和自适应学习功能实现水印信息的盲提取。实验表明,算法对常见的图像处理和几何攻击具有很好的鲁棒性。
This paper presents a blind digital watermarking algorithm based on wavelet transform and neural networks, whose watermark is a binary image and each bit of the watermark is embedded in the multiple positions of the approximate coefficients of the host image through wavelet transform. By distilling the feature points＇ coordinate sets of the approximate coefficients of the watermarked image through wavelet transform, it evaluates the geometrical distortions when a water- marked image suffers from the attack of rotation and zoom, and it corrects the geometrical distortions of the attacked water- marked image using the parameter of geometrical distortions. Finally the extraction of watermarks is implemented by the fine nonlinear mapping and the self-adaptation function of neural networks. The experimental results demonstrate that the proposed scheme is robust against common image processing and geometrical attacks.
出处
《计算机工程与科学》
CSCD
北大核心
2009年第2期40-43,共4页
Computer Engineering & Science
基金
湖北省教育厅科学技术研究项目(D200723001)
关键词
小波变换
几何攻击
特征点坐标集
神经网络
wavelet transform
geometrical attack
feature points coordinate set
neural network
