一种基于Chebyshev混沌神经网络的视频水印算法

A Video Watermarking Algorithm Based on Chebyshev Chaotic Neural Network

针对视频版权的保护,提出一种基于Chebyshev混沌神经网络的视频水印算法。首先选取二值图像作为图像水印,利用Chebyshev混沌神经网络对图像进行加密;然后对宿主视频进行分帧处理,通过Henon映射产生混沌序列提取关键帧,同时提取关键帧亮度分量的低频系数,将处理后的图像水印自适应地嵌入小波变换后亮度分量的低频系数中。实验结果表明,该算法具有较好的不可见性,且针对噪声、高斯滤波、旋转、帧丢失等攻击具有良好的鲁棒性。

For the copyright protection of digital video,this paper proposes a video watermarking algorithm based on Chebyshev chaotic neural network. First,to enhance the watermarking image＇s safety,we process it using Chebyshev chaotic neural network,then decompose the host video into the frames,extract key frames from the chaotic sequences generated by Henon mapping,and extract the low frequence coefficient of the key frame＇s luminance component. The processed image watermark is adaptively embedded into the low frequency coefficients of the luminance component after the wavelet transform. The experimental results show that this algorithm has a good invisibility,and has good robustness for the attacks such as noise,Gauss filtering,rotation,frame shear,etc.