AI Paper Atlas

Research map / Image generation / Super-resolution

OARS: Process-Aware Online Alignment for Generative Real-World Image Super-Resolution

Shijie Zhao, Xuanyu Zhang, Bin Chen, Weiqi Li, Qunliang Xing, Kexin Zhang, Yan Wang, Junlin Li, Li Zhang, Jian Zhang, Tianfan Xue

ECCV 2026 · Image generation › Super-resolution · approach: Perceptual and real-world SR · 2 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 three-stage score calibration (ablation)COMPASS reward · COMPASS-20K reward ablationaccuracy: 78.8 -> 81.5Improved
add explicit fidelity modeling (ablation)COMPASS reward · COMPASS-20K reward ablationaccuracy: 81.5 -> 82.3Improved
add quality-adaptive gamma=7 (ablation)COMPASS reward · COMPASS-20K reward ablationaccuracy: 82.3 -> 83.1Improved
vary gamma among 5, 7, 9 (ablation)COMPASS reward · COMPASS-20K reward ablationaccuracy: 82.7 / 83.1 / 82.8No clear effect
run RL on base model weights instead of SFT weights (stage 1) (ablation)OARS · RealSRLIQE: 4.0784 -> 4.2345Improved
run RL on base model weights instead of SFT weights (both stages) (ablation)OARS · RealSRPSNR: 21.31 -> 22.36Improved
add stage-2 reference-free RL (ablation)OARS on base · RealSRLIQE: 4.2345 -> 4.3045Improved
apply OARS to Qwen-SFTQwen-Image-Edit-2509 · RealSRLIQE and PSNR: LIQE 3.8146 -> 4.3045; PSNR 22.71 -> 22.36Mixed
use COMPASS reward instead of RALI (ablation)OARS RL · RealSRPSNR: 20.83 -> 22.36Improved
add fidelity gating term to perceptual-gain-only reward (ablation)COMPASS reward · RealSRPSNR: 21.38 -> 21.89Improved
add input-quality-aware fidelity term to gated reward (ablation)COMPASS reward · RealSRPSNR: 21.89 -> 22.36Improved
use NFT-based OARS instead of Flow-GRPO (ablation)COMPASS reward · RealSRLIQE: 3.9351 -> 4.3045Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 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