基于DDS和框架的鲁棒盲水印算法  

Robust Blind Watermarking Scheme Based on DDS and Framework

作　　者：朱晓强[1] 洪路宁 周益羽 曾丹[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200444

出　　处：《工业控制计算机》2023年第9期33-35,共3页Industrial Control Computer

基　　金：国家自然科学基金青年基金项目(61402277)。

摘　　要：为进一步提高盲水印算法的鲁棒性,提出了一种基于DWT-DCT-SVD和框架的鲁棒盲水印算法。首先提取彩色宿主图像的Y、U、V分量灰度图,分别进行一级小波变换,对低频子带4×4分块并进行离散余弦变换和奇异值分解;然后通过调节每个子块的对角矩阵,将置乱的水印信息嵌入到宿主图像中;最后在图像四周添加框架用以定位和几何校正。实验结果表明,嵌入水印后图像SSIM和PSNR值分别达到0.96和39.06以上,攻击后提取的水印图成功跳转率在90%以上。该算法在不影响原图视觉效果的前提下,不仅提升了盲水印算法对于旋转、缩放、噪声等常规攻击的鲁棒性,还能抵抗任意角度旋转、大尺度缩放和截图的联合攻击。In order to improve the robustness of the blind watermarking algorithm,a robust blind watermarking algorithm based on DWT-DCT-SVD and framework is proposed in this paper.Firstly,the grayscale images of Y,U,V components of the color host image are extracted,and the first-level wavelet transform is performed,after which the low-frequency sub-band is divided into 4×4 blocks,then discrete cosine transforms and singular value decompositions are performed.By adjusting the diagonal matrix of each sub-block,the scrambled watermark information can be embedded into the host image,and finally a frame around the image is added for positioning and geometric correction.The experimental results show that after the watermark is embedded,the SSIM value and PSNR of the image are above 0.96 and 39.06 respectively,and the successful jump rate of the watermark image extracted after the attack is above 90%.Without affecting the visual effect of the original image,the algorithm not only improves the robustness of the blind watermarking algorithm against conventional attacks such as rotation,scaling,and noise,but also resists joint attacks of arbitrary angle rotation,large-scale zooming,and screenshots.

关 键 词：盲水印 小波变换 奇异值分解 鲁棒性 版权保护 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]