Research map / Image generation / Super-resolution
Efficient Test-Time Adaptation for Super-Resolution with Second-Order Degradation and Reconstruction
Zeshuai Deng, Zhuokun Chen, Shuaicheng Niu, Thomas H. Li, Bohan Zhuang, Mingkui Tan
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 |
|---|---|---|---|
| apply SRTTA test-time adaptation to EDSR | EDSR · DIV2K-C 2x | mean PSNR: 26.21 -> 27.76 | Improved |
| adaptation with consistency loss only (TTA-C) | EDSR · DIV2K-C 2x | mean PSNR: 26.21 -> 26.38 | Improved |
| remove degradation classifier (random second-order degradation) (ablation) | SRTTA · DIV2K-C | mean PSNR: 26.06 -> 27.76 | Improved |
| remove self-supervised loss L_s (ablation) | SRTTA · DIV2K-C | mean PSNR: 27.15 -> 27.76 | Improved |
| remove adaptation consistency loss L_a (ablation) | SRTTA · DIV2K-C | mean PSNR: 10.24 -> 27.76 | Improved |
| alpha below 1 (ablation) | SRTTA · DIV2K-C | mean PSNR: 27.76 -> 22.63 | Hurt |
| alpha above 1 (ablation) | SRTTA · DIV2K-C | mean PSNR: 27.76 -> 27.31 | Hurt |
| increase alpha from 0.5 to 1 (ablation) | SRTTA · Set5 | PSNR: 37.75 -> 34.59 | Hurt |
| adaptive parameter preservation versus random selection (rho=0.5) (ablation) | SRTTA-lifelong · DIV2K-C | mean PSNR: 27.62 -> 27.73 | Improved |
| APP versus stochastic restoration STO (ablation) | SRTTA-lifelong · DIV2K-C | mean PSNR: 27.17 -> 27.73 | Improved |
| STO versus APP on clean images (ablation) | SRTTA-lifelong · Set5 | PSNR: 35.57 -> 34.23 | Hurt |
| lifelong setting without parameter reset | SRTTA · DIV2K-C | mean PSNR: 27.76 -> 27.73 | No clear effect |
Closest papers by meaning
- Unsupervised Diffusion-Based Degradation Modeling for Real-World Super-Resolution (AAAI 2025 · similarity 0.91)
- Unsupervised Degradation Representation Aware Transform for Real-World Blind Image Super-Resolution (AAAI 2025 · similarity 0.90)
- Low-Res Leads the Way: Improving Generalization for Super-Resolution by Self-Supervised Learning (CVPR 2024 · similarity 0.90)
- SinSR: Diffusion-Based Image Super-Resolution in a Single Step (CVPR 2024 · similarity 0.89)
- SEAL: A Framework for Systematic Evaluation of Real-World Super-Resolution (ICLR 2024 · similarity 0.88)
- Towards Realistic Data Generation for Real-World Super-Resolution (ICLR 2025 · 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.
