AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Beyond Blur: A Fluid Perspective on Generative Diffusion Models

Grzegorz Gruszczynski, Jakub Meixner, Michal Wlodarczyk, Przemyslaw Musialski

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Distributional control of diffusion · 1 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add advection with Pe=0.14 to the blur-only forward process (ablation)Inverse heat dissipation (IHD) · FFHQ-128 (sigma=16)FID: 52.10 -> 38.10Improved
add advection with Pe=0.06 to the blur-only forward process (ablation)IHD · FFHQ-128 (sigma=20)FID: 55.87 -> 36.64Improved
increase Peclet number (stronger advection) (ablation)Advection-diffusion model · FFHQ-128 (sigma=16)mean LPIPS (diversity): 0.238 -> 0.302Improved
use very small Pe (0.02 to 0.04) (ablation)Advection-diffusion model · FFHQ-128 (sigma=16)FID: 52.10 -> 52.33 (Pe 0.02), 53.54 (Pe 0.04)No clear effect
raise Pe too high (0.14) at larger final blur (ablation)Advection-diffusion model · FFHQ-128 (sigma=20)FID: 36.64 -> 54.56Hurt
raise Pe from 0.06 to 0.14 (ablation)Advection-diffusion model · FFHQ-128 (sigma=20)Precision: 0.826 -> 0.632Hurt
use larger final blur sigma with same Pe=0 baseline (ablation)IHD · FFHQ-128FID: 52.10 -> 55.87Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy