Research map / Image generation / Super-resolution
From Local Windows to Adaptive Candidates via Individualized Exploratory: Rethinking Attention for Image Super-Resolution
Chunyu Meng, Wei Long, Shuhang Gu
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 propagation (two-hop neighbor exploration) to dilated attention (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.17 -> 26.91 | Improved |
| add propagation to dilated attention (ablation) | IET-light · Manga109 x4 (ablation) | PSNR (dB): 30.58 -> 30.92 | Improved |
| add sparsification to dilated attention (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.17 -> 26.28; FLOPs 180.4G -> 61.2G | Improved |
| add sparsification to propagation (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.91 -> 26.96; FLOPs 215.9G -> 68.9G | Improved |
| add SF-FFN on top of propagation and sparsification (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.96 -> 27.03 | Improved |
| add SF-FFN on top of propagation and sparsification (ablation) | IET-light · Manga109 x4 (ablation) | PSNR (dB): 31.05 -> 31.09 | Improved |
| increase propagation steps from 1 to 4 (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.52 -> 27.03 | Improved |
| increase propagation steps from 4 to 5 (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 27.03 -> 26.96 | No clear effect |
| increase dilation from 1 to 3 for train and inference (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.86 -> 26.99 | Improved |
| train with d=2 and infer with d=3 instead of d=3 for both (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.99 -> 27.03 | Improved |
| set k1/k2 ratio to 2 instead of 0.5 (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.98 -> 27.03 | Improved |
| place SF-FFN after the last two blocks vs none (ablation) | IET-light · Urban100 x4 (IET-light ablation) | PSNR (dB): 26.96 -> 27.03 | Improved |
Closest papers by meaning
- Transcending the Limit of Local Window: Advanced Super-Resolution Transformer with Adaptive Token Dictionary (CVPR 2024 · similarity 0.90)
- Recursive Generalization Transformer for Image Super-Resolution (ICLR 2024 · similarity 0.89)
- CFAT: Unleashing Triangular Windows for Image Super-resolution (CVPR 2024 · similarity 0.88)
- Efficient Attention-Sharing Information Distillation Transformer for Lightweight Single Image Super-Resolution (AAAI 2025 · similarity 0.88)
- Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment (NeurIPS 2025 · similarity 0.88)
- AdaFormer: Efficient Transformer with Adaptive Token Sparsification for Image Super-resolution (AAAI 2024 · 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.
