摘要
提出了基于神经模糊控制网络的新的自适应图像水印算法。该算法首先对原始图像进行小波变换,组织小波块。接着,从子块的视觉掩蔽性和能量强度两个方面,建立和训练神经模糊控制网络,得到每个小波子块嵌入水印的最优强度值。最后,根据神经模糊控制得出的最优嵌入强度,将不同强度的水印嵌入到不同的小波块中。实验结果表明,所提出的算法能抵抗各种常规的图像处理的攻击,具有较好的鲁棒性和透明性。
A novel adaptive watermarking algorithm was proposed for image based on neural-fuzzy control. Firstly the original image was decomposed by wavelet transformation. And then some wavelet coefficient were organized to form wavelet blocks. After that, the neural-fuzzy control network was set up and trained in order to get the best embedding strength of the watermark. Finally, watermark components with different strength were embedded to different wavelet blocks according to the best strength got fi＇om the trained network. The experimental results show that the proposed algorithm is with good robustness and imperceptibility to several image processing and attacks.
出处
《计算机工程与设计》
CSCD
北大核心
2005年第10期2657-2659,2697,共4页
Computer Engineering and Design
基金
福建省教育厅科技基金项目(K99029)
关键词
数字水印
小波包变换
自适应
神经网络
模糊控制
digital watermarking
wavelet packet transformation
adaptive
neural network
fuzzy control
