基于ICA和DWT的自适应盲水印算法

为了使数字水印具有较强的鲁棒性和不可见性,提出一种基于独立分量分析（ICA）和离散小波变换（DWT）的自适应鲁棒性盲水印算法.该算法先将载体图像进行小波变换,然后将原始水印进行平铺和置乱,再根据噪声可见函数（NVF）的视觉掩蔽作用,将其自适应地嵌入到载体图像小波变换域的低频系数上,最后利用ICA技术实现了水印盲提取.实验结果表明,该算法的水印不可见性较好,且可抵抗多种常见水印攻击,其中对JPEG压缩、尺度缩放等攻击的鲁棒性较突出,可应用于图像的版权保护.

Adaptive Blind Watermarking Algorithm Based on ICA and DWT

In order to enhance the robustness and invisibility of digital watermarking,a new adaptive watermarking algorithm based on independent component analysis(ICA) and discrete wavelet transform(DWT) is proposed.The algorithm firstly decomposes the image into wavelet domain.Then the watermark is adaptive embedded into its low frequency coefficient,under the visual mask of noise visibility function(NVF),which has been rearranged by tiling extended and scramble before embedded.It realizes blind extraction by ICA.Experimental results show that the algorithm has good transparency and better robustness,especially in JPEG compression and resizing scale attacks.It can be applicable for digital image copyright protection.