Research map / Image generation / Super-resolution
FRAMER: Frequency-Aligned Self-Distillation with Adaptive Modulation Leveraging Diffusion Priors for Real-World Image Super-Resolution
Seungho Choi, Jeahun Sung, Jihyong Oh
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 frequency-specific contrastive self-distillation instead of MSE distillation (ablation) | DiT4SR (RealSR) · RealSR | MANIQA: 0.467 -> 0.533 | Improved |
| add frequency decomposition to MSE distillation (ablation) | DiT4SR (RealSR) · RealSR | MUSIQ: 68.11 -> 65.74 | Hurt |
| use final layer as teacher instead of a random layer (ablation) | FRAMER · RealSR | MANIQA: 0.469 -> 0.533 | Improved |
| use random-layer negative instead of previous-layer negative (ablation) | FRAMER · RealSR | PSNR: 22.23 -> 23.04 | Improved |
| add InterCL for HF alongside IntraCL for LF (ablation) | FRAMER · RealSR | MUSIQ: 65.92 -> 69.01 | Improved |
| assign InterCL to LF and IntraCL to HF (reversed) (ablation) | FRAMER · RealSR | MANIQA: 0.533 -> 0.484 | Hurt |
| add FAW (adaptive weights) (ablation) | FRAMER contrastive distillation · RealSR | MUSIQ: 69.01 -> 70.50 | Improved |
| add FAM (alignment gating) (ablation) | FRAMER contrastive distillation · RealSR | PSNR: 23.04 -> 23.06 | No clear effect |
| add both FAW and FAM (ablation) | FRAMER contrastive distillation · RealSR | MUSIQ: 69.01 -> 72.24 | Improved |
| apply FRAMER to DiT4SR (DiT) | DiT4SR · DrealSR | PSNR: 23.64 -> 24.73 | Improved |
| apply FRAMER to PiSA-SR (U-Net) | PiSA-SR · RealSR | NIQE: 5.902 -> 5.513 | Improved |
| apply FRAMER to PiSA-SR (U-Net) | PiSA-SR · RealLR200 | NIQE: 4.410 -> 4.381 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DiT4SR — DiT4SR: Taming Diffusion Transformer for Real-World Image Super-Resolution (ICCV 2025) · better on 4 of 4 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
- Noise-Started One-Step Image Super-Resolution via LR-Conditioned SplitMeanFlow and GAN Refinement (NeurIPS 2026 · similarity 0.90)
- Adversarial Diffusion Compression for Real-World Image Super-Resolution (CVPR 2025 · similarity 0.90)
- Selective Diffusion Distillation for Real-World High-Scale Image Super-Resolution (AAAI 2026 · similarity 0.89)
- NSARM: Next-Scale Autoregressive Modeling for Robust Real-World Image Super-Resolution (NeurIPS 2026 · similarity 0.88)
- Uncertainty-guided Perturbation for Image Super-Resolution Diffusion Model (CVPR 2025 · similarity 0.88)
- When Latents Forget Pixels: Restoring Fidelity in Diffusion Transformer Super-Resolution (NeurIPS 2026 · similarity 0.88)
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.
