AI Paper Atlas

Research map / Image generation / Image restoration

FoundIR-v2: Optimizing Pre-Training Data Mixtures for Image Restoration Foundation Model

Xiang Chen, Jinshan Pan, Jiangxin Dong, Jian Yang, Jinhui Tang

CVPR 2026 · Image generation › Image restoration · approach: Efficient restoration architectures · 5 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
add data equilibrium scheduling (DES) instead of mixed training (ablation)FoundIR-v2 · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 18.91 -> 20.41Improved
add DES instead of sequential learning (ablation)FoundIR-v2 · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 18.69 -> 20.41Improved
add DES instead of incremental learning (ablation)FoundIR-v2 · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 19.93 -> 20.41Improved
add DES to PromptIR (ablation)PromptIR · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 15.89 -> 16.92Improved
add DES to FoundIR (ablation)FoundIR · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 16.77 -> 17.95Improved
replace FoundIR model with FoundIR-v2 under DES (no SR task) (ablation)FoundIR-v2 · Four-task avg (deblur, dehaze, low-light, SR)PSNR: 17.95 -> 20.09Improved
use equal or inappropriate data mixture proportions (ablation)all-in-one restoration model · Deblur/dehaze/low-light/SRPSNR per taskHurt
vary evaluation interval T (10000, 25000, 30000) (ablation)DES · Four-task avg (deblur, dehaze, low-light, SR) Avg PSNRAvg. PSNR: 20.17 / 20.53 / 20.41No clear effect
vary D_ref size (10, 20, 30) (ablation)DES · Four-task avg (deblur, dehaze, low-light, SR) Avg PSNRAvg. PSNR: 20.41 / 20.48 / 20.36No clear effect
add soft MoE-driven scheduler to SDXL (ablation)FoundIR-v2 · Four sub-tasks (radar chart)task PSNR/MUSIQImproved
filter low-quality GT images from training data (ablation)FoundIR-v2 · Four-task averageMUSIQImproved
use FoundIR-v2 versus best prior on UHD-LL and FoundIR-LFoundIR-v2 · UHD-LL low-lightPSNR: 22.52 (AutoDIR) vs 20.08Hurt

Later papers that compared with this method

1 later accepted paper reports FoundIR-v2 as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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