Image Watermarking of Generative Diffusion Models

2025-02-12

AI Summary

To address copyright protection challenges in generative diffusion models, this paper proposes an end-to-end differentiable framework that implicitly embeds watermarks into the model’s parameter space—departing from conventional pixel- or frequency-domain watermarking paradigms. Methodologically, it jointly optimizes a watermarked diffusion generator and a dedicated watermark extractor, leveraging implicit feature modeling and adversarial reconstruction to ensure high invisibility, diversity of watermark embedding, and robust extraction. Key contributions include: (i) the first approach to intrinsically integrate watermarks into the diffusion model architecture itself, enabling cross-model unique identification and precise provenance tracing; and (ii) superior performance—achieving >99.2% watermark detection accuracy under diverse attacks (e.g., compression, noise, fine-tuning), while maintaining visual imperceptibility, strong generalization across architectures, and discriminative capability among distinct models.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Data transparency and provenance
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Embedding watermarks into the output of generative models is essential for establishing copyright and verifiable ownership over the generated content. Emerging diffusion model watermarking methods either embed watermarks in the frequency domain or offer limited versatility of the watermark patterns in the image space, which allows simplistic detection and removal of the watermarks from the generated content. To address this issue, we propose a watermarking technique that embeds watermark features into the diffusion model itself. Our technique enables training of a paired watermark extractor for a generative model that is learned through an end-to-end process. The extractor forces the generator, during training, to effectively embed versatile, imperceptible watermarks in the generated content while simultaneously ensuring their precise recovery. We demonstrate highly accurate watermark embedding/detection and show that it is also possible to distinguish between different watermarks embedded with our method to differentiate between generative models.

Problem

• Embedding watermarks in diffusion models
• Ensuring versatile, imperceptible watermark patterns
• Enabling precise recovery of embedded watermarks

Innovation

• Embed watermark in diffusion model
• Train paired watermark extractor end-to-end
• Enable versatile imperceptible watermark recovery