Research map / Image generation / Super-resolution
Beyond the Ground Truth: Enhanced Supervision for Image Restoration
Donghun Ryou, Inju Ha, Sanghyeok Chu, Bohyung Han
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add ORNet on top of FFTformer | FFTformer · GoPro deblurring | MUSIQ: 46.47 -> 64.57 | Improved |
| add ORNet on top of AdaRevD | AdaRevD · GoPro deblurring | A-FINE: 36.19 -> 14.38 | Improved |
| add ORNet on top of NAFNet | NAFNet · SIDD denoising | MUSIQ: 22.73 -> 35.87 | Improved |
| add ORNet on top of FFTformer under extra Gaussian blur | FFTformer · GoPro OOD (extra Gaussian blur) | MUSIQ: 22.3812 -> 42.9131 | Improved |
| attach ORNet instead of finetuning with enhanced supervision (ablation) | FFTformer · GoPro deblurring | MUSIQ: 60.54 -> 64.57 | Improved |
| finetune FFTformer on enhanced ground truth (ablation) | FFTformer · GoPro deblurring | MUSIQ: 46.47 -> 60.54 | Improved |
| use mixed 2x, 3x, 4x GT variants instead of only the 4x variant (ablation) | ORNet · GoPro deblurring | MUSIQ: 62.81 -> 64.57 | Improved |
| increase enhancement level lambda from 0.1 to 0.9 (ablation) | ORNet on FFTformer · GoPro deblurring | MUSIQ: 48.07 -> 69.76 | Improved |
| increase lambda from 0.3 to 0.9 (ablation) | ORNet on FFTformer · GoPro deblurring | VisualQuality-R1: 4.1995 -> 3.9600 | Hurt |
| increase lambda from 0.3 to 0.9 under extra Gaussian blur (ablation) | ORNet on FFTformer · GoPro OOD (extra Gaussian blur) | PSNR (original GT): 24.5774 -> 24.2407 | Hurt |
| use 25 Gaussian ring bases instead of 5 (ablation) | frequency mask generator · GoPro deblurring (qualitative) | frequency control: B=5 -> B=25 | Improved |
| refine with ORNet instead of applying the OSEDiff ISR model directly (ablation) | AdaRevD · GoPro deblurring (qualitative) | hallucination | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- ResShift — ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting (NeurIPS 2023) · better on 1 of 1 shared result
- Retinexformer — Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement (ICCV 2023) · better on 1 of 1 shared result
- Restormer — Restormer: Efficient Transformer for High-Resolution Image Restoration (CVPR 2022) · better on 1 of 1 shared result
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
- Low-Res Leads the Way: Improving Generalization for Super-Resolution by Self-Supervised Learning (CVPR 2024 · similarity 0.87)
- From Zero to Detail: Deconstructing Ultra-High-Definition Image Restoration from Progressive Spectral Perspective (CVPR 2025 · similarity 0.87)
- Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild (CVPR 2024 · similarity 0.87)
- Boosting Image Restoration via Priors from Pre-trained Models (CVPR 2024 · similarity 0.86)
- ReFIR: Grounding Large Restoration Models with Retrieval Augmentation (NeurIPS 2024 · similarity 0.86)
- Continuous Expert Assembly: Instance-Conditioned Low-Rank Residuals for All-in-One Image Restoration (NeurIPS 2026 · similarity 0.86)
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.
