基于SVD和小波包分解的自适应鲁棒水印算法

针对当前水印鲁棒性和透明性矛盾的问题,提出了一种基于奇异值变换(SVD)和小波包分解的自适应鲁棒水印算法,对二值水印图像进行Arnold置乱预处理,以增强水印信息的安全性。在水印嵌入过程中,将原始宿主图像分为8×8子图像块,并对每一子块进行小波包分解。根据人眼视觉特性和图像块自身的亮度以及纹理特征确定最佳量化步长,将水印信息通过量化调制的方法自适应地嵌入至相应高低频区域的奇异值中。实验结果表明,该算法在具有良好透明性的同时,对JPEG压缩、加噪、滤波、几何攻击等常见的攻击方式也有着较强的鲁棒性。

Adaptive robust watermarking algorithm based onSVD and wavelet packet transform

Aiming at the problem of contradiction between roubustness and invisibility of the watermark, this paper proposed a new adaptive robust watermarking algorithm based on SVD and wavelet packet decomposition. The algorithm made pretreatment of binary watermark image with Arnold scrambling to enhance security. In watermarking embedding process, it firstly divided the host image into 8×8 sub-blocks and decomposed each block by wavelet packet. It calculated the optimal quantization step according to HVS(human visual system) and the block's brightness and texture features. Finally, embedded the watermark in the singular values of corresponding high and low frequency coefficients adaptively by quantization index modulation. The experimental results show that the proposed algorithim can obtain better visibilty and it is robust against attacks such as JPEG compression added noise, and filtering.