AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Tuning Real-World Image Restoration at Inference: A Test-Time Scaling Paradigm for Flow Matching Models

Purui Bai, Junxian Duan, Pin Wang, Jinhua Hao, Ming Sun, Chao Zhou, Huaibo Huang

ECCV 2026 · Image generation › Fast diffusion sampling · approach: Flow matching distillation · 1 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 VLM text caption condition (ablation)ResFlow-Tuner · LSDIR-ValLPIPS: 0.3856 -> 0.3721Improved
use UMMF unified-sequence fusion instead of direct adding (ablation)ResFlow-Tuner · LSDIR-ValFID: 28.39 -> 19.82Improved
use UMMF instead of ControlNet (ablation)ResFlow-Tuner · LSDIR-ValFID: 24.73 -> 19.82Improved
use three-reward verifier ensemble instead of Aesthetic alone (ablation)ResFlow-Tuner · LSDIR-ValMANIQA: 0.6224 -> 0.6658Improved
use verifier ensemble instead of CLIPScore + ImageReward pair (ablation)ResFlow-Tuner · LSDIR-ValMUSIQ: 70.98 -> 74.21Improved
add test-time scaling (K=4, N=7) (ablation)ResFlow-Tuner · LSDIR-ValFID: 22.91 -> 19.82Improved
increase TTS to K=10, N=15 (ablation)ResFlow-Tuner · LSDIR-ValFID: 19.82 -> 18.17Improved
increase TTS to K=10, N=15 (compute cost) (ablation)ResFlow-Tuner · LSDIR-Valinference compute: 0.375e4 -> 1.0e4Hurt
use ResFlow-Tuner versus GAN baselines on full-reference fidelityResFlow-Tuner · DIV2K-Val-IPSNR: 27.63 -> 23.11Hurt
use ResFlow-Tuner versus FaithDiff on no-reference qualityResFlow-Tuner · DIV2K-Val-IMUSIQ: 66.53 -> 75.54Improved
restore with ResFlow-Tuner before OCRResFlow-Tuner · Occluded RoadText 2024OCR precision (%): 36.45% -> 42.23%Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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