[2507.18534] Elucidating the Design Space of Arbitrary-Noise-Based Diffusion Models

Computer Science > Computer Vision and Pattern Recognition arXiv:2507.18534 (cs) [Submitted on 24 Jul 2025 (v1), last revised 5 Mar 2026 (this version, v2)] Title:Elucidating the Design Space of Arbitrary-Noise-Based Diffusion Models Authors:Xingyu Qiu, Mengying Yang, Xinghua Ma, Dong Liang, Fanding Li, Gongning Luo, Wei Wang, Kuanquan Wang, Shuo Li View a PDF of the paper titled Elucidating the Design Space of Arbitrary-Noise-Based Diffusion Models, by Xingyu Qiu and 8 other authors View PDF HTML (experimental) Abstract:Although EDM aims to unify the design space of diffusion models, its reliance on fixed Gaussian noise prevents it from explaining emerging flow-based methods that diffuse arbitrary noise. Moreover, our study reveals that EDM's forcible injection of Gaussian noise has adverse effects on image restoration task, as it corrupts the degraded images, overextends the restoration distance, and increases the task's complexity. To interpret diverse methods for handling distinct noise patterns within a unified theoretical framework and to minimize the restoration distance, we propose EDA, which Elucidates the Design space of Arbitrary-noise diffusion models. Theoretically, EDA expands noise pattern flexibility while preserving EDM's modularity, with rigorous proof that increased noise complexity introduces no additional computational overhead during restoration. EDA is validated on three representative medical image denoising and natural image restoration tasks: MRI ...