GenPTW: In-Generation Image Watermarking for Provenance Tracing and Tamper Localization

2025-04-28

AI Summary

To address the dual challenges of AIGC authenticity verification and copyright protection, existing image watermarking methods struggle to simultaneously achieve high fidelity, robustness, and precise tampering localization—often relying on post-processing or reference images. This paper proposes, for the first time, the endogenous embedding of structured watermarks directly into the diffusion generation process. Our method builds upon Latent Diffusion Models (LDMs) to construct an end-to-end watermarking architecture featuring a novel frequency-domain collaborative decoder and an AIGC-editing distortion simulation layer, explicitly modeling domain-specific generative distortions. Experiments demonstrate that our approach surpasses state-of-the-art methods across all key metrics: visual fidelity, watermark extraction accuracy, and fine-grained tampering localization. Notably, it achieves significantly enhanced robustness under complex post-editing scenarios, enabling reliable provenance tracing and localized integrity verification for generated images.

Technology Category

Computer Vision: Diffusion Models for VisionMachine Learning: Large Multimodal Models (LMMs)Natural Language Processing: Generation

Application Category

Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAISearch and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAGWeb Mining and Content Analysis: Web data provenance, reliability, and authenticity

Abstract

The rapid development of generative image models has brought tremendous opportunities to AI-generated content (AIGC) creation, while also introducing critical challenges in ensuring content authenticity and copyright ownership. Existing image watermarking methods, though partially effective, often rely on post-processing or reference images, and struggle to balance fidelity, robustness, and tamper localization. To address these limitations, we propose GenPTW, an In-Generation image watermarking framework for latent diffusion models (LDMs), which integrates Provenance Tracing and Tamper Localization into a unified Watermark-based design. It embeds structured watermark signals during the image generation phase, enabling unified provenance tracing and tamper localization. For extraction, we construct a frequency-coordinated decoder to improve robustness and localization precision in complex editing scenarios. Additionally, a distortion layer that simulates AIGC editing is introduced to enhance robustness. Extensive experiments demonstrate that GenPTW outperforms existing methods in image fidelity, watermark extraction accuracy, and tamper localization performance, offering an efficient and practical solution for trustworthy AIGC image generation.

Problem

• Ensuring authenticity and copyright in AI-generated images
• Balancing fidelity, robustness, and tamper localization in watermarks
• Integrating provenance tracing and tamper detection during image generation

Innovation

• In-Generation watermarking for latent diffusion models
• Frequency-coordinated decoder for robust extraction
• Distortion layer simulating AIGC editing enhances robustness