Research map / Image generation / Super-resolution
FreqOrtho-SR: Frequency-Guided Orthogonal Expert Learning for Real-World Image Super-Resolution
Minh Son Hoang, Dinh Tran, Quyen Nguyen Duc, Phuong Dam, Daeyoung Kim
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-guided routing to MoE gating (ablation) | FreqMoE · RealSR | FID: 118.35 -> 111.45 | Improved |
| Add frequency-guided routing to MoE gating (ablation) | FreqMoE · RealSR | PSNR: 26.3718 -> 26.3097 | Hurt |
| Add multi-expert MoE with spatial gating (ablation) | single LoRA dual-LoRA · RealSR | PSNR: +0.87 dB | Improved |
| Add OGP to FreqMoE (ablation) | FreqMoE · RealSR | FID: 111.45 -> 108.91 | Improved |
| Add OGP to FreqMoE (ablation) | FreqMoE · RealSR | PSNR: 26.3097 -> 26.27 | Hurt |
| Add OGP without multi-expert routing (ablation) | dual-LoRA · RealSR | FID: 113.39 (OGP only) vs 118.35 (MoE only) | Improved |
| Add OGP (ablation) | semantic LoRA · per-layer subspace overlap | projection energy ratio: 87% -> near zero (output layer) | Improved |
| Replace PiSA-SR with FreqOrtho-SR | one-step SR · DRealSR | PSNR: 28.31 -> 29.08 | Improved |
| Replace PiSA-SR with FreqOrtho-SR | one-step SR · RealSR | MUSIQ: 70.15 -> 69.39 | Hurt |
| Replace PiSA-SR with FreqOrtho-SR | one-step SR · RealSR | inference time (s): 0.09 -> 0.30 | Hurt |
| Replace PiSA-SR with FreqOrtho-SR | one-step SR · RealLR200 | NIQE: 4.00 -> 3.90 | Improved |
Closest papers by meaning
- Pixel-level and Semantic-level Adjustable Super-resolution: A Dual-LoRA Approach (CVPR 2025 · similarity 0.89)
- Frequency Consistent Adaptation for Real World Super Resolution (AAAI 2021 · similarity 0.89)
- IFCSR: Inference-Free Fidelity-Realism Control for One-Step Diffusion-based Real-World Image Super-Resolution (CVPR 2026 · similarity 0.88)
- Mixture of Ranks with Degradation-Aware Routing for One-Step Real-World Image Super-Resolution (AAAI 2026 · similarity 0.88)
- Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator (NeurIPS 2025 · similarity 0.86)
- NSARM: Next-Scale Autoregressive Modeling for Robust Real-World Image Super-Resolution (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.
