DiffuseTrace: A Transparent and Flexible Watermarking Scheme for Latent Diffusion Model

TL;DR

DiffuseTrace is a novel watermarking scheme for latent diffusion models that embeds multi-bit watermarks semantically without affecting image quality, ensuring robustness against various attacks and easy integration as a plug-in.

Contribution

It introduces a flexible, non-fine-tuning watermarking method for diffusion models that maintains high detection accuracy and robustness against attacks.

Findings

Maintains 99% watermark detection rate under attack

Achieves over 94% attribution accuracy

Does not compromise image quality

Abstract

Latent Diffusion Models (LDMs) enable a wide range of applications but raise ethical concerns regarding illegal utilization. Adding watermarks to generative model outputs is a vital technique employed for copyright tracking and mitigating potential risks associated with Artificial Intelligence (AI)-generated contents. However, post-processed watermarking methods are unable to withstand generative watermark attacks and there exists a trade-off between image fidelity and watermark strength. Therefore, we propose a novel technique called DiffuseTrace. DiffuseTrace does not rely on fine-tuning of the diffusion model components. The multi-bit watermark is a embedded into the image space semantically without compromising image quality. The watermark component can be utilized as a plug-in in arbitrary diffusion models. We validate through experiments the effectiveness and flexibility of DiffuseTrace. Under 8 types of image processing watermark attacks and 3 types of generative watermark attacks, DiffuseTrace maintains watermark detection rate of 99% and attribution accuracy of over 94%.