基于生成对抗网络的图像鲁棒水印改进模型

doi:10.12066/j.issn.1007-2861.2496

摘要/Abstract

摘要： 目前,水印算法所生成的水印图像存在对不可微噪声处理的鲁棒性较低、图像质量一般的缺陷,介绍了一种基于生成对抗网络的图像鲁棒水印改进模型.该模型将空间和通道注意力机制模块融合到具有密集结构类型的编码器中,增强其特征提取能力;在鉴别器前端增加高通滤波器,使生成的水印图像具有较好的不可感知性;在噪声层中使用一种小批量真实和模拟(mini batch of real and simulated,MBRS) JPEG处理,且添加各种不同强度类型的噪声进行端到端训练.研究结果表明,相比HiDDeN与二阶段深度学习鲁棒水印模型,该模型生成的水印图像具有较高的鲁棒性和不可感知性,平均峰值信噪比(peak signal to noise ratio,PSNR)提高了1.20 dB,平均结构相似性提高了4.71%,平均信息提取误码率降低了13.55%.

关键词: 生成对抗网络, 鲁棒水印, 注意力机制, 特征提取, 不可感知性

Abstract: The watermark image generated by the current watermarking algorithm exhibits poor robustness to non-differentiable noise processing and inadequate general image quality. This study introduces an improved model for a robust image watermarking based on generative adversarial networks. The model integrates the spatial and channel attention mechanism blocks into an encoder with dense structure type to enhance feature extraction ability. A high-pass filter is added to the front of the discriminator to enhance the imperceptibility of the generated watermark image. For end-to-end training, mini-batch of real and simulated (MBRS) JPEG processing is performed in the noise layer, adding various types of noise of different strengths. The experimental results show that compared with HiDDeN and two-stage deep learning robust watermarking model, the watermarked images generated by this model have better robustness and imperceptibility. The average peak signal-to-noise ratio (PSNR) increased by 1.20 dB; average structural similarity increased by 4.71%; and average information extraction bit error rate was reduced by 13.55%.

Key words: generative adversarial networks, robust watermarking, attention mechanism, feature extraction, imperceptibility

中图分类号:

TP309.7

赵亚宁, 严利民. 基于生成对抗网络的图像鲁棒水印改进模型[J]. 上海大学学报(自然科学版), 2026, 32(2): 226-239.

ZHAO Yaning, YAN Limin. Improved model for robust image watermarking based on generative adversarial networks[J]. Journal of Shanghai University（Natural Science Edition）, 2026, 32(2): 226-239.

参考文献

[1] 张卫明, 王宏霞, 李斌, 等. 多媒体隐写研究进展[J]. 中国图象图形学报, 2022, 27(6): 1918-1943.
[2] 徐伟, 刘颖, 朱婷鸽. 基于DCT和SVD的图像哈希水印算法[J]. 计算机工程与设计, 2020, 41(1): 145-149.
[3] 童人婷, 程航, 张新鹏. 基于EVD变换的鲁棒音频水印算法[J]. 上海大学学报(自然科学版), 2016, 22(4): 388-397.
[4] Kandi H, Mishra D, Gorthi S R K S. Exploring the learning capabilities of convolutional neural networks for robust image watermarking [J]. Computers & Security, 2017, 65: 247-268
[5] Zhu J, Kaplan R, Johnson J, et al. Hidden: hiding data with deep networks [C]// Proceedings of the European Conference on Computer Vision (ECCV). 2018: 657-672.
[6] Goodfellow I J, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets [J]. Advances in Neural Information Processing Systems. 2014: 27.
[7] Liu Y, Guo M X, Zhang J, et al. A novel two-stage separable deep learning framework for practical blind watermarking [C]// Proceedings of the 27th ACM International Conference on Multimedia. 2019: 1509-1517.
[8] He K M, Zhang X Y, Ren S Q, et al. Deep residual learning for image recognition [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016: 770-778.
[9] 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.
[10] Zhang K A, Cuesta-Infante A, Xu L, et al. SteganoGAN: high capacity image steganography with GANs [J]. Computer Vision and Pattern Recognition. 2019. DOI: 10.48550/arXiv.1901.03892.
[11] 王照乾, 孔韦韦, 滕金保, 等. DenseNet生成对抗网络低照度图像增强方法[J]. 计算机工程与应用, 2022, 58(8): 214-220.
[12] Sharma S S, Chandrasekaran V. A robust hybrid digital watermarking technique against a powerful CNN-based adversarial attack [J]. Multimedia Tools and Applications, 2020, 79(43): 32769-32790.
[13] Hao K, Feng G, Zhang X. Robust image watermarking based on generative adversarial network [J]. China Communications, 2020, 17(11): 131-140.
[14] 郑钢, 胡东辉, 戈辉, 等. 生成对抗网络驱动的图像隐写与水印模型[J]. 中国图象图形学报, 2021, 26(10): 2485-2502.
[15] Liu K L, Chen D D, Liao J, et al. JPEG robust invertible grayscale [J]. IEEE Transactions on Visualization and Computer Graphics. 2021. DOI: 10.1109/TVCG.2021.3088531.
[16] Jia Z Y, Fang H, Zhang W M. MBRS: enhancing robustness of dnn-based watermarking by mini-batch of real and simulated JPEG compression [C]// Proceedings of the 29th ACM International Conference on Multimedia. 2021: 41-49.
[17] Byrnes O, La W, Wang H, et al. Data hiding with deep learning: a survey unifying digital watermarking and steganography [J]. IEEE Transactions on Computational Social Systems, 2023, 10(6): 2985-2999.