基于Hopfield的DWT数字图像水印新算法

为了保证数字图像水印的鲁棒性和安全性,结合Hopfield神经网络的联想记忆功能,本文提出了一种基于Hopfield神经网络的离散小波变换域数字图像水印新算法。该算法首先对数字图像进行二尺度的离散小波变换,然后再使用Hopfield神经网络对提取的水印进行多级处理和恢复,从而明显地提高了水印图像提取的质量,增强了水印算法的鲁棒性。仿真实验表明:该算法能很好地抵抗噪声污染J、PEG压缩和剪切等攻击。

A novel DWT digital image watermarking based on Hopfield

A new digital image watermark algorithm of discrete wavelet transform was proposed to guarantee its robustness and security by combining with association memory ability of Hopfield neural network.First,the 2-level discrete wavelet transform was applied to the digital image,then Hopfield neural network disposed of and got back the extracted watermark repeatedly,so it could improve the extracted watermarking image to strength the robustness of the algorithm.Experimental results showed that this method could an resist against noise adding,JPEG compression,image cut and so on.