AI Paper Atlas

Research map / Image generation / Super-resolution

One-Step Residual Shifting Diffusion for Image Super-Resolution via Distillation

Daniil Selikhanovych, David Li, Aleksei Leonov, Nikita Gushchin, Sergei Kushneriuk, Alexander Filippov, Evgeny Burnaev, Iaroslav Sergeevich Koshelev, Alexander Korotin

ICML 2026 · Image generation › Super-resolution · approach: Diffusion-based super-resolution

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
train RSD with N=4 timesteps instead of single-step N=1 (ablation)RSD · RealSRLPIPS: 0.4052 -> 0.3552Improved
increase timesteps N from 4 to 15 (ablation)RSD · RealSRPSNR / CLIPIQA: PSNR 24.92 -> 25.91; CLIPIQA 0.7518 -> 0.6857Mixed
increase timesteps N from 1 to 8 (ablation)RSD · RealSRMUSIQ: 64.290 -> 66.445Improved
add both supervised losses (lambda_1, lambda_2 nonzero) to RSD (ablation)RSD · RealSRPSNR / LPIPS: PSNR 24.92 -> 25.91; LPIPS 0.3552 -> 0.2726Improved
add both supervised losses to RSD (ablation)RSD · RealSRCLIPIQA / MUSIQ: CLIPIQA 0.7518 -> 0.7060; MUSIQ 66.430 -> 65.860Hurt
enable only lambda_2 supervised loss (ablation)RSD · RealSRMUSIQ: 66.430 -> 67.615Improved
distill ResShift with RSD loss instead of SinSRResShift teacher · RealSRLPIPS: 0.365 -> 0.273Improved
use RSD loss instead of VSD loss for ResShift distillationResShift one-step student · RealSRLPIPS / PSNR: LPIPS 0.466 -> 0.273; PSNR 23.96 -> 25.91Improved
RSD versus SinSR training/inference efficiencyRSD · ComplexityGPU memory / training time: 539 MB vs 570 MB; 5 / 4 A100 vs 60 / 1 A100 hoursImproved
compare RSD one-step with OSEDiff on DIV2K-ValRSD · DIV2K-ValPSNR: 23.72 -> 23.91 (RSD) but vs CTMSR 24.88Hurt

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