基于混沌的Contourlet域抗几何攻击图像水印算法

为了提高水印算法的抗几何攻击性和自适应性,提出了一种基于混沌的Contourlet域抗几何攻击的自适应图像水印算法。首先对载体图像进行Contourlet变换,并对低频子带进行分块奇异值分解,然后根据HVS原理自适应地将混沌加密后的水印信息嵌入到每块的最大奇异值中;最后使用改进的Harris-Laplace算子提取含水印图像的特征点。水印提取时利用受攻击前后图像的特征点矫正图像的几何变形。实验表明,该算法对JPEG压缩、滤波、加噪、旋转、缩放和平移等常规的图像处理和几何攻击具有很强的鲁棒性。

Image watermarking algorithm against geometrical attacks inContourlet domain based on chaos

In order to improve the robustness against geometrical attacks and self-adaptability, this paper proposed an adaptive image watermarking algorithm in Contourlet domain based on chaos. Firstly, it applied a Contourlet transform to the cover image, and divided the low frequency sub-band into many blocks which were decomposed with singular value decompositionSVD. Secondly, it adaptively embedded the watermark information with chaotic encryption into the largest singular value of every block based on human visual system HVS. Finally, it extracted the feature points by the improved Harris-Laplace operator. During detection, restored the attacked imageby using feature points of it and those of the watermarked image. Experimental results show the proposed algorithm is robust to common image processing and geometric attacks such as JPEG compression, filtering, noise addition, rotation, scaling and translation.