基于PCA和DWT的强鲁棒数字水印算法

针对传统离散小波变换(DWT)数字水印算法抗几何攻击能力较弱的问题,提出一种基于主成分分析(PCA)和DWT的新的数字水印算法。新算法对载体图像进行一级小波分解,在低频子带上用主成分分析提取出既含有高频又含有低频成分的主成分系数,将水印嵌入到提取出的主成分系数中。实验结果表明,与传统DWT水印算法相比,该算法不仅明显提高了抗剪裁、旋转等抗几何攻击能力,对加噪、图像灰度值变化等攻击也表现出了很强的鲁棒性。

Robust digital watermarking algorithm based on PCA and DWT

The digital watermarking algorithm based on discrete wavelet transform (DWT) is weakly resistant to geometric. This paper presents a digital watermarking algorithm based on principal component analysis and DWT transform. In this algorithm, the host image is first transformed with DWT, and principle component analysis (PCA) is used to de-correlate the image pixel to obtain the principle components. The watermark is then embedded into the principle component. Compared with the watermarking algorithm based on DWT, experimental results indicate that the new algorithm shows a strong robustness:it not only greatly improves the ability against geometric attacks such as anti-cropping, rotation and others, but also has good resistance to noise, change of image gray value and other attacks.