DiffusionShield: A Watermark for Copyright Protection against Generative Diffusion Models

TL;DR

Generative Diffusion Models (GDMs) raise copyright concerns by reproducing works without permission. The paper presents DiffusionShield, a watermarking framework that uses pattern uniformity and blockwise patch encoding to embed ownership messages into images, ensuring the watermark reappears in GDM outputs. A bi-level joint optimization of watermark design and a decoding detector yields high detection accuracy with minimal distortion and scales to multiple owners and long messages. Extensive experiments across datasets show strong protection performance, robustness to fine-tuning and corruptions, and practical feasibility for copyright enforcement in diffusion-based generation scenarios.

Abstract

Recently, Generative Diffusion Models (GDMs) have showcased their remarkable capabilities in learning and generating images. A large community of GDMs has naturally emerged, further promoting the diversified applications of GDMs in various fields. However, this unrestricted proliferation has raised serious concerns about copyright protection. For example, artists including painters and photographers are becoming increasingly concerned that GDMs could effortlessly replicate their unique creative works without authorization. In response to these challenges, we introduce a novel watermarking scheme, DiffusionShield, tailored for GDMs. DiffusionShield protects images from copyright infringement by GDMs through encoding the ownership information into an imperceptible watermark and injecting it into the images. Its watermark can be easily learned by GDMs and will be reproduced in their generated images. By detecting the watermark from generated images, copyright infringement can be exposed with evidence. Benefiting from the uniformity of the watermarks and the joint optimization method, DiffusionShield ensures low distortion of the original image, high watermark detection performance, and the ability to embed lengthy messages. We conduct rigorous and comprehensive experiments to show the effectiveness of DiffusionShield in defending against infringement by GDMs and its superiority over traditional watermarking methods. The code for DiffusionShield is accessible in https://github.com/Yingqiancui/DiffusionShield.

Figures (17)

• Figure 1: Watermark detection accuracy (%) on GDM-generated images and the corresponding budget ($l_2$ norm) of watermarks.
• Figure 2: An overview of watermarking with two stages.
• Figure 3: Uniformity vs. watermark detection rate.
• Figure 4: An $8\times8$ sequence of basic patches encoded with message "103313131232...". Different patterns represent different basic patches.
• Figure 5: Average bit accuracy (%) across different numbers of copyright owners (on class-conditional GDM).

Theorems & Definitions (3)

• proof : Proof of Example \ref{['example:linear']}
• Remark 2.1
• proof : Proof of Example \ref{['example:mlp']}