计算机科学 ›› 2024, Vol. 51 ›› Issue (2): 172-181.doi: 10.11896/jsjkx.230600144
赵江锋1, 和红杰1, 陈帆2, 杨树斌1
ZHAO Jiangfeng1, HE Hongjie1, CHEN Fan2, YANG Shubin1
摘要: 可见水印是一种常用的数字图像版权保护手段。分析可见水印去除结果可以验证图像上水印的有效性,并为水印设计者提供设计或添加水印的参考和启发。目前,大多数的水印去除方法都是基于自然图像的研究,而文档图像在生活中也被广泛使用,但由于缺乏公开的文档图像去水印数据集,相关文档图像的水印去除研究较少。为了探究水印去除方法在文档图像上的水印去除效果,构建了一个文档图像水印去除数据集(SDIWRD)。在对文档图像可见水印去除的研究中发现,使用已有的水印去除方法得到的水印去除结果中容易留下水印主体伪影或者轮廓伪影。为了解决这个问题,提出了一种基于全局与局部特征的二阶段文档图像可见水印去除模型(RWRNet),该模型采用由粗到细的二阶段的半实例归一化编解码器架构。在粗略阶段,使用全局与局部特征提取模块增强对全局空间特征的捕捉能力,同时保留对局部细节信息的提取能力,从而帮助进行水印去除;在细化阶段,细化网络共享粗略阶段权重,并构建循环特征融合模块来充分挖掘粗略阶段编解码器的重要特征,为细化阶段提供丰富的上下文信息,帮助进行细致的水印去除。此外,还结合了结构相似性损失来帮助获取更好的视觉质量。所提方法在SDIWRD数据集上进行了实验,实验结果显示PSNR达到了41.21 dB,SSIM达到了99.07%,RMSE降低至3.64,优于现有水印去除方法。另外也在公开的CLWD彩色水印去除数据集进行了实验,实验结果显示PNSR达到了39.31 dB,SSIM达到98.81%,RMSE降低至3.50,也优于现有水印去除方法。实验结果证明了所提方法具有良好的泛化性和去水印的能力,能有效减轻水印伪影。最后还提出了一些防止水印去除的建议,在相关网站1)可公开访问所提出的方法和数据集。
中图分类号:
[1]HUANG C H,WU J L.Attacking visible watermarkingschemes [J].IEEE Transactions on Multimedia,2004,6(1):16-30. [2]JAESIK P,TAI Y W,KWEON I S.Identigram/Watermark removal using cross-channel correlation[C]//2012 IEEE Confe-rence on Computer Vision & Pattern Recognition.IEEE,2012:446-453. [3]PEI S C,ZENG Y C.A novel image recovery algorithm for visible watermarked images[J].IEEE Transactions on Information Forensics & Security,2006,1(4):543-550. [4]DEKEL T,RUBINSTEIN M,LIU C,et al.On the effectiveness of visible watermarks[C]//Proceedings of the IEEE Conference on Computer Vision & Pattern Recognition.2017:2146-2154. [5]SHARMA M,VERMA A,VIG L.Learning to clean:A GANperspective[C]//Computer Vision-ACCV 2018 Workshops:14th Asian Conference on Computer Vision.Perth,Australia:Springer International Publishing,2019:174-185. [6]CAO Z,NIU S,ZHANG J,et al.Generative adversarial net-works model for visible watermark removal[J].IET Image Processing,2019,13(10):1783-1789. [7]LI X,LU C,CHENG D,et al.Towards photo-realistic visiblewatermark removal with conditional generative adversarial networks[C]//Image & Graphics:10th International Conference.Beijing,China:Springer International Publishing,2019:345-356. [8]WANG J L,LIU X Q,LI B Y,et al.A Scheme of Visible Watermark Removal Method Based on Conditional Generative Adversarial Nets[J].Computer Technology and Development,2022,32(2):119-124. [9]CHENG D,LI X,LI W H,et al.Large-scale visible watermark detection and removal with deep convolutional networks[C]//Pattern Recognition & Computer Vision:First Chinese Confe-rence.Guangzhou,China:Springer International Publishing,2018:27-40. [10]HERTZ A,FOGEL S,HANOCKA R,et al.Blind visual motifremoval from a single image[C]//Proceedings of the IEEE/CVF Conference on Computer Vision & Pattern Recognition.2019:6858-6867. [11]CUN X,PUN C M.Split then refine:stacked attention-guidedResUNets for blind single image visible watermark removal[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2021,35(2):1184-1192. [12]LIU Y,ZHU Z,BAI X.Wdnet:Watermark-decomposition net-work for visible watermark removal[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision.2021:3685-3693. [13]LIANG J,NIU L,GUO F,et al.Visible watermark removal via self-calibrated localization and background refinement[C]//Proceedings of the 29th ACM International Conference on Multimedia.2021:4426-4434. [14]RONNEBERGER O,FISCHER P,BROX T.U-net:Convolu-tional networks for biomedical image segmentation[C]//18th International Conference Medical Image Computing & Compu-ter-Assisted Intervention(MICCAI 2015).Munich,Germany:Springer International Publishing,2015:234-241. [15]ZHANG M M,ZHOU Q,HU Y L.Visible watermark removal scheme based on multiple matching[J].Computer Engineering and Design,2020,41(1):176-182. [16]ZHU J Y,PARK T,ISOLA P,et al.Unpaired image-to-image translation using cycle-consistent adversarial networks[C]//Proceedings of the IEEE International Conference on Computer Vision.2017:2223-2232. [17]YU F,KOLTUN V.Multi-scale context aggregation by dilated convolutions[EB/OL].(2015-11-23) [2023-06-11].https://arxiv.org/abs/1511.07122. [18]DOSOVITSKIY A,BEYER L,KOLESNIKOV A,et al.Animage is worth 16x16 words:Transformers for image recognition at scale[EB/OL].(2020-10-22)[2023-06-11].https://arxiv.org/abs/2010.11929. [19]LIU Z,LIN Y,CAO Y,et al.Swin transformer:Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.2021:10012-10022. [20]ZAMIR S W,ARORA A,KHAN S,et al.Multi-stage progressive image restoration[C]//Proceedings of the IEEE/CVF Conference on Computer Vision & Pattern Recognition.2021:14821-14831. [21]FAN X N,ZHAO Z X,YAN W,et al.Multi-scale Feature Fusion Image Dehazing Algorithm Combined with Attention Mecha-nism[J].Computer Science,2022,49(5):50-57. [22]CHEN L,LU X,ZHANG J,et al.Hinet:Half instance normalization network for image restoration[C]//Proceedings of the IEEE/CVF Conference on Computer Vision & Pattern Recognition.2021:182-192. [23]WOO S,PARK J,LEE J Y,et al.Cbam:Convolutional block attention module[C]//Proceedings of the European Conference on Computer Vision(ECCV).2018:3-19. [24]MA Y,YU D,WU T,et al.PaddlePaddle:An open-source deep learning platform from industrial practice[J].Frontiers of Data &Computing,2019,1(1):105-115. [25]KINGMA D P,BA J.Adam:A method for stochastic optimization[EB/OL].(2014-12-22) [2023-06-11].https://arxiv.org/abs/1412.6980. |
|