Research map / Image generation / Super-resolution
RRSR:Reciprocal Reference-Based Image Super-Resolution with Progressive Feature Alignment and Selection
Lin Zhang, Xin Li, Dongliang He, Fu Li, Yili Wang, Zhaoxiang Zhang
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 progressive feature alignment (PFA) (ablation) | C2-Matching baseline · CUFED5 | PSNR: 28.40 -> 28.63 | Improved |
| add reference-aware feature selection (RAS) (ablation) | Baseline+PFA · CUFED5 | PSNR: 28.63 -> 28.70 | Improved |
| add reciprocal target-reference reconstruction (RTRR) training (ablation) | Baseline+PFA+RAS · CUFED5 | PSNR: 28.70 -> 28.83 | Improved |
| add RTRR training (ablation) | C2-Matching · CUFED5 | PSNR: 28.40 -> 28.64 | Improved |
| add RTRR training (ablation) | MASA · CUFED5 | PSNR: 27.47 -> 27.58 | Improved |
| add RTRR training (ablation) | TTSR · CUFED5 | PSNR: 27.03 -> 27.17 | Improved |
| increase lambda_RTRR beyond 0.4 (ablation) | RRSR with RTRR · CUFED5 | PSNR | Hurt |
| use learnable filter count 8 or 32 instead of 16 in RAS (ablation) | RRSR · CUFED5 | PSNR: decrease by around 0.02 dB | No clear effect |
| apply random perspective transform to reference LR/HR in second pass (ablation) | RTRR · CUFED5 | reconstruction quality | Improved |
| use RRSR with RTRR vs C2-Matching | RRSR · CUFED5 | PSNR: 27.16 -> 28.09 | Improved |
| use RRSR on Urban100 vs NLSN | RRSR-rec · Urban100 | PSNR: 26.28 -> 26.21 | Hurt |
| use RRSR-rec vs C2-Matching-rec on Sun80 | RRSR-rec · Sun80 | PSNR: 30.18 -> 30.13 | No clear effect |
Closest papers by meaning
- Robust Reference-Based Super-Resolution With Similarity-Aware Deformable Convolution (CVPR 2020 · similarity 0.91)
- Image Super-Resolution by Neural Texture Transfer (CVPR 2019 · similarity 0.90)
- Robust Reference-Based Super-Resolution via C2-Matching (CVPR 2021 · similarity 0.90)
- Reference-Based Image Super-Resolution with Deformable Attention Transformer (ECCV 2022 · similarity 0.90)
- MASA-SR: Matching Acceleration and Spatial Adaptation for Reference-Based Image Super-Resolution (CVPR 2021 · similarity 0.90)
- Building Bridges across Spatial and Temporal Resolutions: Reference-Based Super-Resolution via Change Priors and Conditional Diffusion Model (CVPR 2024 · similarity 0.85)
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.
