Research map / Image generation / Super-resolution
MASA-SR: Matching Acceleration and Spatial Adaptation for Reference-Based Image Super-Resolution
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use MASA-rec rather than TTSR-rec | RefSR · CUFED5 | PSNR: 27.09 -> 27.54 | Helped |
| use MASA-rec rather than TTSR-rec | RefSR · Sun80 | PSNR: 30.02 -> 30.15 | Helped |
| use MASA-rec rather than TTSR-rec | RefSR · Urban100 | PSNR: 25.87 -> 26.09 | Helped |
| use coarse-to-fine matching rather than dense full-map matching | MASA · 128×128 LR, 512×512 Ref | matching FLOPS (G): 618.48 -> 8.84 | Helped |
| increase LR block size (ablation) | MASA matching · CUFED5 testing set | PSNR | Hurt |
| increase LR block size (ablation) | MASA matching · CUFED5 testing set | FLOPS | Helped |
| use LR block size 1×1 (ablation) | MASA matching · CUFED5 testing set | PSNR: 27.60 dB | Helped |
| use LR block size 1×1 (ablation) | MASA matching · CUFED5 testing set | FLOPS: 787.77G | Hurt |
| increase Ref↓ block size (ablation) | MASA matching · CUFED5 testing set | PSNR | Helped |
| increase Ref↓ block size (ablation) | MASA matching · CUFED5 testing set | FLOPS | Hurt |
| use more combinations of dilation rates (ablation) | MASA matching · CUFED5 testing set | PSNR | Helped |
| add SAM instead of using no normalization (ablation) | MASA · CUFED5 testing set | PSNR: 27.32 -> 27.54 (+0.22 dB) | Helped |
Closest papers by meaning
- Robust Reference-Based Super-Resolution via C2-Matching (CVPR 2021 · similarity 0.92)
- Task Decoupled Framework for Reference-Based Super-Resolution (CVPR 2022 · similarity 0.91)
- Image Super-Resolution by Neural Texture Transfer (CVPR 2019 · similarity 0.91)
- Building Bridges across Spatial and Temporal Resolutions: Reference-Based Super-Resolution via Change Priors and Conditional Diffusion Model (CVPR 2024 · similarity 0.87)
- KeDuSR: Real-World Dual-Lens Super-Resolution via Kernel-Free Matching (AAAI 2024 · similarity 0.84)
- IODA: Instance-Guided One-shot Domain Adaptation for Super-Resolution (NeurIPS 2024 · similarity 0.83)
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.