References & Citations
Computer Science > Computer Vision and Pattern Recognition
Title: Detecting Image Attribution for Text-to-Image Diffusion Models in RGB and Beyond
(Submitted on 28 Mar 2024 (v1), last revised 10 Apr 2024 (this version, v2))
Abstract: Modern text-to-image (T2I) diffusion models can generate images with remarkable realism and creativity. These advancements have sparked research in fake image detection and attribution, yet prior studies have not fully explored the practical and scientific dimensions of this task. In addition to attributing images to 12 state-of-the-art T2I generators, we provide extensive analyses on what inference stage hyperparameters and image modifications are discernible. Our experiments reveal that initialization seeds are highly detectable, along with other subtle variations in the image generation process to some extent. We further investigate what visual traces are leveraged in image attribution by perturbing high-frequency details and employing mid-level representations of image style and structure. Notably, altering high-frequency information causes only slight reductions in accuracy, and training an attributor on style representations outperforms training on RGB images. Our analyses underscore that fake images are detectable and attributable at various levels of visual granularity than previously explored.
Submission history
From: Katherine Xu [view email][v1] Thu, 28 Mar 2024 17:59:42 GMT (26801kb,D)
[v2] Wed, 10 Apr 2024 20:03:05 GMT (26802kb,D)
Link back to: arXiv, form interface, contact.