输出方式对印刷品全息水印提取质量的影响

目的为实现对印刷输出图像中提取水印质量的有效预估,研究硬拷贝输出方式对印刷品全息防伪水印提取质量的影响。方法先将傅里叶全息水印的高频分量嵌入宿主图像经2次小波变换的中频区域,实现信息隐藏。使用喷墨打印、激光打印、胶版印刷、数码印刷等4种常用方式对含水印防伪图像进行硬拷贝输出,然后从承印物纸张表面质量、图像色彩空间转换和输出精度等3个方面,利用NC参数对硬拷贝图像的水印提取质量进行评价,并对比分析影响水印提取质量的因素。结果数字图像、打印图像、印刷图像间的水印提取质量有较大的差别,总体表现为喷墨打印胶版印刷数码印刷和激光打印。承印物基材较高的白度、平滑度有利于水印提取。在输出与输入过程中,含水印图像经历的多次色彩空间转换对原始图像的色彩分布、复现质量有一定程度的改变,进而影响水印提取质量。在满足图像复现质量要求的前提下,输出精度对水印提取质量的影响较小。结论研究水印的抗打印/扫描鲁棒性时,不能用模拟攻击与桌面打印代替实际输出方式。选择表面质量高的承印物有利于防伪水印信息的识别。验证水印方案的抗打印/扫描鲁棒性时,其分辨率满足图像的复现质量要求即可。

Influence of Output Types on Extraction Quality of Holographic Watermark

The work aims to study the influence of hardcopy types on extraction quality of holographic anti-counterfeiting watermark in the printed image, in order to effectively predict the quality of watermark extracted from the images outputted from the printing. Firstly, the high-frequency component of Fourier holographic watermark was embedded into the middle-frequency area of the host image subject to wavelet transform for twice, so as to hide the information. Then, four common hardcopy types (inkjet, laser, offset printing and digital printing) were used to output watermarked anti-counterfeiting images, and the watermark extraction quality of hardcopy image was evaluated with NC parameters from three aspects, including paper surface quality of printed material, color space conversion of images and output accuracy. In addition, the factors that influence the watermark extraction quality were compared and analyzed. Experiment results showed that the watermark extraction qualities between digital image, presswork and printing images were very different, and the rank was: inkjet printing> offset printing> digital printing and laser printing. The printed substrate paper with higher whiteness and gloss was beneficial to watermark extraction. The multiple color space conver-sions of watermarked image in output and input process to some extent affected the color distribution and reproduction quality of original image and further the watermark extraction quality. On the premises of satisfying the image reproduction quality, the output resolution had little effect on watermark extraction quality. It is not possible to replace the actual output type with simulated attacks or desktop printing when studying the anti-print/scan robustness of watermarking. The substrate paper with high surface quality facilitates the identification of anti-counterfeiting watermark information. The resolution is appropriate as long as the reproduction quality requirements are satisfied when verifying the anti-print/scan robustness of watermarking scheme.