面向AIGC生成图像的水印发展进程
The Development Progress of Watermarking for AIGC-Generated Images
摘要: 随着人工智能生成内容(AIGC)技术的迅速发展，扩散模型(Diffusion Models, DMs)在图像生成和编辑领域展现了巨大的潜力。然而，扩散模型生成内容的真实性难以辨别，滥用生成模型可能引发隐私泄露、版权侵权等社会问题。为了解决这一问题，数字水印技术逐渐成为保护AIGC模型版权、追踪生成内容的重要手段。本文从图像生成技术的发展、传统和最新的数字图像水印算法以及针对AIGC的水印方法三个核心领域进行了综述。此外，我们还研究了该领域中使用的常见性能评估指标。最后，对研究中存在的问题进行了讨论，并提出了今后的研究方向。
Abstract: With the rapid advancement of Artificial Intelligence Generated Content (AIGC) technologies, diffusion models (DMs) have demonstrated tremendous potential in image generation and editing. However, it is often difficult to distinguish the authenticity of content generated by diffusion models, and their misuse may lead to issues such as privacy breaches and copyright infringement. To address these challenges, digital watermarking has emerged as a crucial approach for protecting the copyright of AIGC models and tracing generated content. This paper reviews the development of image generation technologies, traditional and cutting-edge digital watermarking algorithms, and watermarking methods specifically designed for AIGC. Furthermore, commonly used performance evaluation metrics in this field are examined. Finally, the paper discusses current research challenges and outlines future research directions.
文章引用：耿新晨. 面向AIGC生成图像的水印发展进程[J]. 人工智能与机器人研究, 2025, 14(4): 967-976. https://doi.org/10.12677/airr.2025.144092

参考文献

[1] Van den Oord, A., Kalchbrenner, N., Espeholt, L., et al. (2016) Conditional Image Generation with PixelCNN Decoders. Advances in Neural Information Processing Systems, 29.
[2] Kingma, D.P. and Welling, M. (2013) Auto-Encoding Variational Bayes. [Google Scholar] [CrossRef] 
[3] Goodfellow, I.J., Pouget-Abadie, J., Mirza, M., et al. (2014) Generative Adversarial Nets. Advances in Neural Information Processing Systems, 27.
[4] Sohl-Dickstein, J., Weiss, E., Maheswaranathan, N., et al. (2015) Deep Unsupervised Learning Using Nonequilibrium Thermodynamics. International Conference on Machine Learning, Lille, 6-11 July 2015, 2256-2265.
[5] Ho, J., Jain, A. and Abbeel, P. (2020) Denoising Diffusion Probabilistic Models. Advances in Neural Information Processing Systems, 33, 6840-6851.
[6] Song, J., Meng, C. and Ermon, S. (2020) Denoising Diffusion Implicit Models. arXiv: 2010.02502.
[7] Rombach, R., Blattmann, A., Lorenz, D., Esser, P. and Ommer, B. (2022) High-Resolution Image Synthesis with Latent Diffusion Models. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, 18-24 June 2022, 10674-10685. [Google Scholar] [CrossRef] 
[8] 王树梅. 数字图像水印技术综述[J]. 湖南理工学院学报(自然科学版), 2022, 35(1): 31-36+68.
[9] 冯乐, 朱仁杰, 吴汉舟, 等. 神经网络水印综述[J]. 应用科学学报, 2021, 39(6): 881-892.
[10] 赵蕾, 桂小林, 邵屹杨, 等. 数字图像多功能水印综述[J]. 计算机辅助设计与图形学学报, 2024, 36(2): 195-222.
[11] Barni, M., Bartolini, F., Cappellini, V. and Piva, A. (1998) A DCT-Domain System for Robust Image Watermarking. Signal Processing, 66, 357-372. [Google Scholar] [CrossRef] 
[12] Piva, A., Barni, M., Bartolini, F. and Cappellini, V. (1997) DCT-Based Watermark Recovering without Resorting to the Uncorrupted Original Image. Proceedings of International Conference on Image Processing, Santa Barbara, 26-29 October 1997, 520-523. [Google Scholar] [CrossRef] 
[13] Daren, H., Jiufen, L., Jiwu, H., et al. (2001) A DWT-Based Image Watermarking Algorithm. IEEE International Conference on Multimedia and Expo, 2001. Tokyo, 22-25 August 2001, 313-316.
[14] Jiansheng, M., Sukang, L. and Xiaomei, T. (2009) A Digital Watermarking Algorithm Based on DCT and DWT. The 2009 International Symposium on Web Information Systems and Applications (WISA 2009), Academy Publisher, 104.
[15] Urvoy, M., Goudia, D. and Autrusseau, F. (2014) Perceptual DFT Watermarking with Improved Detection and Robustness to Geometrical Distortions. IEEE Transactions on Information Forensics and Security, 9, 1108-1119. [Google Scholar] [CrossRef] 
[16] Kang, X., Huang, J., Shi, Y.Q., et al. (2003) A DWT-DFT Composite Watermarking Scheme Robust to Both Affine Transform and JPEG Compression. IEEE Transactions on Circuits and Systems for Video Technology, 13, 776-786.
[17] 夏道勋, 王林娜, 宋允飞, 等. 深度神经网络模型数字水印技术研究进展综述[J]. 科学技术与工程, 2023, 23(5): 1799-1811.
[18] Zhu, J. (2018) HiDDeN: Hiding Data with Deep Networks. arXiv: 1807. 09937.
[19] Cox, I., Miller, M., Bloom, J., et al. (2007) Digital Watermarking and Steganography. Morgan Kaufmann.
[20] Zhang, K.A., Xu, L., Cuesta-Infante, A., et al. (2019) Robust Invisible Video Watermarking with Attention. arXiv: 1909. 01285.
[21] Cui, Y., Ren, J., Xu, H., et al. (2023) Diffusionshield: A Watermark for Copyright Protection against Generative Diffusion Models. arXiv: 2306. 04642.
[22] Fernandez, P., Couairon, G., Jégou, H., Douze, M. and Furon, T. (2023) The Stable Signature: Rooting Watermarks in Latent Diffusion Models. 2023 IEEE/CVF International Conference on Computer Vision (ICCV), Paris, 1-6 October 2023, 22409-22420. [Google Scholar] [CrossRef] 
[23] Liu, Y., Li, Z., Backes, M., et al. (2023) Watermarking Diffusion Model. arXiv: 2305. 12502.
[24] Xiong, C., Qin, C., Feng, G. and Zhang, X. (2023) Flexible and Secure Watermarking for Latent Diffusion Model. Proceedings of the 31st ACM International Conference on Multimedia, Ottawa, 29 October 2023-3 November 2023, 1668-1676. [Google Scholar] [CrossRef] 
[25] Wen, Y., Kirchenbauer, J., Geiping, J., et al. (2023) Tree-Ring Watermarks: Fingerprints for Diffusion Images that Are Invisible and Robust. arXiv: 2305. 20030.

为你推荐