摘要
目的数字印刷和手机拍摄是检验含有数字水印的印刷产品版权有效性的主要途径,解决在数字印刷和手机拍摄场景下水印信息难以提取这一问题。方法提出一种基于离散小波变换和奇异值分解的抗数字印刷/手机拍摄的数字水印算法。首先对宿主图像进行二级离散小波变换,对低频子带进行奇异值分解,将Arnold置乱后的水印嵌入到奇异值分解的对角矩阵中,最后进行小波逆变换得到嵌入水印图像。结果含水印信息的峰值信噪比(PSNR)达到了39.67,在不同攻击下的归一化相关系数(NC)均在0.89以上,实际的印刷拍摄测试中NC值也均达到0.94,证明该水印算法在保证不可见性的同时对于印刷产品实际应用中的常见攻击具有很好的抵抗能力。结论对不同品牌的智能手机、印刷设备以及不同种类的纸张采集到的含水印图像进行测试,均能够在印刷和拍摄场景下实现水印提取,证明了该算法的有效性。
Digital printing and mobile shooting are the main ways to verify the validity of copyright of printed products containing digital watermarks. Therefore, the work aims to solve the problem that it is difficult to extract watermark in digital printing and mobile shooting. A digital watermarking algorithm based on discrete wavelet transform and singular value decomposition to resist digital printing/mobile shooting was proposed. Firstly, the host image was processed by two-stage discrete wavelet transform, and singular value decomposition was performed on the low-frequency sub-bands.The watermark after Arnold was scrambled was embedded into the diagonal matrix of singular value decomposition. Finally, the embedded watermark image was obtained by inverse wavelet transform. The peak signal to noise ratio(PSNR)containing watermark information reached 39.67. The normalized correlation coefficient(NC) under different attacks was above 0.89. The NC value in actual shooting test also reached 0.94, showing that watermark algorithm not only ensured invisibility but also had good resistance to common attacks in practical application of printing products. The watermark images collected from different brands of smart phones, printing devices and different kinds of paper can be extracted in printing and shooting scenes, which proves the effectiveness of the algorithm.
作者
张明贵
高静
ZHANG Ming-gui;GAO Jing(Sichuan Kuanzhai Printing Co.,Ltd.,Deyang 618100,China;The Great Wall Cigar Factory,China Tobacco Sichuan Industrial Co.,Ltd.,Shifang 618400,China)
出处
《包装工程》
CAS
北大核心
2022年第3期274-283,共10页
Packaging Engineering
关键词
数字水印
打印拍摄
离散小波变换
奇异值分解
digital watermarking
printing and shooting
discrete wavelet transform
singular value decomposition