Towards Robust Content Watermarking Against Removal and Forgery Attacks

2026-04-08

AI Summary

This work addresses the vulnerability of existing text-to-image diffusion models to watermark removal and forgery attacks, which undermines copyright protection and content provenance. To overcome these limitations, the authors propose ISTS (Instance-Specific watermarking with Two-Sided detection), a novel paradigm that integrates instance-specific watermark generation with a two-sided detection mechanism. The method dynamically controls the timing and pattern of watermark embedding based on the semantic content of user prompts and employs a two-stage detection algorithm. Experimental results demonstrate that ISTS significantly outperforms current state-of-the-art approaches in resisting both removal and forgery attacks, achieving markedly improved watermark detection accuracy and robustness.

Abstract

Generated contents have raised serious concerns about copyright protection, image provenance, and credit attribution. A potential solution for these problems is watermarking. Recently, content watermarking for text-to-image diffusion models has been studied extensively for its effective detection utility and robustness. However, these watermarking techniques are vulnerable to potential adversarial attacks, such as removal attacks and forgery attacks. In this paper, we build a novel watermarking paradigm called Instance-Specific watermarking with Two-Sided detection (ISTS) to resist removal and forgery attacks. Specifically, we introduce a strategy that dynamically controls the injection time and watermarking patterns based on the semantics of users' prompts. Furthermore, we propose a new two-sided detection approach to enhance robustness in watermark detection. Experiments have demonstrated the superiority of our watermarking against removal and forgery attacks.

Problem

• content watermarking
• removal attacks
• forgery attacks
• diffusion models
• copyright protection

Innovation

• instance-specific watermarking
• two-sided detection
• removal attack
• forgery attack
• diffusion models