ImageSentinel: Protecting Visual Datasets from Unauthorized Retrieval-Augmented Image Generation

2025-10-13

AI Summary

In retrieval-augmented image generation (RAIG), private visual datasets face unauthorized usage risks, and conventional digital watermarking fails due to feature recombination during generation. Method: We propose a novel copyright protection framework that leverages vision-language models to synthesize sentinel images—visually consistent with the original data yet implicitly encoding dataset provenance—and couples them with random character sequences as retrieval keys, enabling fine-grained usage tracking and verifiable attribution. The method requires no modification to the RAIG backbone, ensuring compatibility with existing systems while preserving watermark robustness and generated image fidelity. Results: Experiments demonstrate high-precision detection of data misuse; authorized generations incur negligible quality degradation (ΔFID < 0.3). The implementation is publicly available.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Safety and Robustness

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The widespread adoption of Retrieval-Augmented Image Generation (RAIG) has raised significant concerns about the unauthorized use of private image datasets. While these systems have shown remarkable capabilities in enhancing generation quality through reference images, protecting visual datasets from unauthorized use in such systems remains a challenging problem. Traditional digital watermarking approaches face limitations in RAIG systems, as the complex feature extraction and recombination processes fail to preserve watermark signals during generation. To address these challenges, we propose ImageSentinel, a novel framework for protecting visual datasets in RAIG. Our framework synthesizes sentinel images that maintain visual consistency with the original dataset. These sentinels enable protection verification through randomly generated character sequences that serve as retrieval keys. To ensure seamless integration, we leverage vision-language models to generate the sentinel images. Experimental results demonstrate that ImageSentinel effectively detects unauthorized dataset usage while preserving generation quality for authorized applications. Code is available at https://github.com/luo-ziyuan/ImageSentinel.

Problem

• Protecting visual datasets from unauthorized retrieval-augmented image generation systems
• Overcoming traditional watermarking limitations in complex feature extraction processes
• Enabling protection verification while maintaining dataset visual consistency

Innovation

• Synthesizes sentinel images for dataset protection
• Uses character sequences as retrieval keys for verification
• Leverages vision-language models to generate sentinels