Research map / Image generation / Super-resolution
ATD: Improved Transformer With Adaptive Token Dictionary for Image Restoration
Leheng Zhang, Wei Long, Yawei Li, Xingyu Zhou, Xiaorui Zhao, 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 token dictionary cross-attention (TDCA) to baseline (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.40 -> 26.46 | Improved |
| add token dictionary cross-attention (TDCA) to baseline (ablation) | ATD-light · Manga109 x4 (ATD-light) | PSNR: 30.84 -> 30.92 | Improved |
| add adaptive category-based self-attention (ACMSA) after TDCA (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.46 -> 26.70 | Improved |
| add adaptive category-based self-attention (ACMSA) after TDCA (ablation) | ATD-light · Manga109 x4 (ATD-light) | PSNR: 30.92 -> 31.19 | Improved |
| add category-aware FFN (CFFN) (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.70 -> 26.76 | Improved |
| add category-aware FFN (CFFN) (ablation) | ATD-light · Manga109 x4 (ATD-light) | PSNR: 31.19 -> 31.21 | Improved |
| use reparameterized scaling factor tau' at dictionary size 128 (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.70 -> 26.76 | Improved |
| increase dictionary size M from 64 to 128 with original scaling factor (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.73 -> 26.70 | Hurt |
| increase sub-category size n_s from 0 (no ACMSA) to 128 (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.46 -> 26.76 | Improved |
| increase sub-category size n_s from 128 to 256 (ablation) | ATD-light · Urban100 x4 (ATD-light) | PSNR: 26.76 -> 26.75 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MambaIRv2 — MambaIRv2: Attentive State Space Restoration (CVPR 2025) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Transcending the Limit of Local Window: Advanced Super-Resolution Transformer with Adaptive Token Dictionary (CVPR 2024 · similarity 0.89)
- Adapt or Perish: Adaptive Sparse Transformer with Attentive Feature Refinement for Image Restoration (CVPR 2024 · similarity 0.89)
- AdaFormer: Efficient Transformer with Adaptive Token Sparsification for Image Super-resolution (AAAI 2024 · similarity 0.88)
- Sharing Key Semantics in Transformer Makes Efficient Image Restoration (NeurIPS 2024 · similarity 0.88)
- Correlation Matching Transformation Transformers for UHD Image Restoration (AAAI 2024 · similarity 0.87)
- Continuous Expert Assembly: Instance-Conditioned Low-Rank Residuals for All-in-One Image Restoration (NeurIPS 2026 · 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.
