Detection Limits and Statistical Separability of Tree Ring Watermarks in Rectified Flow-based Text-to-Image Generation Models
This work addresses the problem of authenticating AI-generated images for users of advanced text-to-image models, but it is incremental as it evaluates an existing technique on new models.
The paper investigated the effectiveness of Tree-Ring Watermarking in rectified flow-based text-to-image models like SD 2.1 and FLUX.1-dev, finding that noise latent inversion limitations reduce watermark recovery and statistical separability, with specific detection rates dropping under certain conditions.
Tree-Ring Watermarking is a significant technique for authenticating AI-generated images. However, its effectiveness in rectified flow-based models remains unexplored, particularly given the inherent challenges of these models with noise latent inversion. Through extensive experimentation, we evaluated and compared the detection and separability of watermarks between SD 2.1 and FLUX.1-dev models. By analyzing various text guidance configurations and augmentation attacks, we demonstrate how inversion limitations affect both watermark recovery and the statistical separation between watermarked and unwatermarked images. Our findings provide valuable insights into the current limitations of Tree-Ring Watermarking in the current SOTA models and highlight the critical need for improved inversion methods to achieve reliable watermark detection and separability. The official implementation, dataset release and all experimental results are available at this \href{https://github.com/dsgiitr/flux-watermarking}{\textbf{link}}.