AI Paper Atlas

Research map / Image generation / Super-resolution

Training-free, Perceptually Consistent Low-Resolution Previews with High-Resolution Image for Efficient Workflows of Diffusion Models

Wongi Jeong, Hoigi Seo, Se Young Chun

CVPR 2026 · Image generation › Super-resolution · approach: Diffusion-based super-resolution · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
select downsample matrix by minimum commutator norm instead of nearest downsampling (ablation)Preview generation · FLUX.1-dev, 500 promptsDreamSim: 10.39 -> 7.05Improved
select downsample matrix by minimum commutator norm instead of random (ablation)Preview generation · FLUX.1-dev, 500 promptsPSNR (dB): 20.851 -> 20.962Improved
select by argmin instead of argmax of commutator (ablation)Preview generation · FLUX.1-dev, 500 promptsDreamSim: 8.55 -> 7.05Improved
add commutator-zero guidance (ablation)Preview generation · FLUX.1-dev, 500 promptsDreamSim: 8.56 -> 7.05Improved
add commutator-zero guidance (ablation)Preview generation · FLUX.1-dev, 500 promptsPSNR (dB): 19.115 -> 20.962Improved
increase correction iterations m from 3 to 7 (ablation)commutator-zero guidance · FLUX.1-dev PSNR sweepPSNR (dB): 20.645 -> 21.389 at alpha=0.04Improved
increase step size alpha beyond optimum (ablation)commutator-zero guidance · FLUX.1-dev PSNR sweepPSNR (dB): 21.146 -> 20.543 (m=5, alpha 0.04 -> 0.06)Hurt
increase m (more correction steps) (ablation)commutator-zero guidance · compute costcomputationHurt
combine with TaylorSeer temporal-axis accelerationPreview generation · FLUX.1-devDreamSim: 16.97 -> 7.79 vs FLUX.1-dev (10) at similar speedupImproved
downsample at timestep tD with commutator-zero condition instead of naive downsamplingPreview generation · FLUX.1-devDreamSim: 9.20 -> 6.83Improved
downsample at timestep tD instead of generating directly at low resolutionPreview generation · FLUX.1-devDreamSim: 21.02 -> 6.83Improved
downsample latents instead of reducing NFE to 20Preview generation · SD3.5-LPIQE: 29.39 ref; ours 31.55 vs naive 29.75Hurt

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