Research map / Architectures and efficiency / Attention and transformers
Cross Aggregation Transformer for Image Restoration
Read the paperCheck 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 |
|---|---|---|---|
| replace square window with rectangle window self-attention (ablation) | window Transformer for SR · Urban100 x2 | PSNR: 32.50 -> 32.66 | Helped |
| add axial-shift to rectangle-window attention (ablation) | Rwin-SA · Urban100 x2 | PSNR: 32.66 -> 32.91 | Helped |
| add swin shift to square window attention (ablation) | square window attention · Urban100 x2 | PSNR: 32.50 -> 32.75 | Helped |
| add locality complementary module (LCM) (ablation) | CAT-A · Urban100 x2 | PSNR: 33.01 -> 33.11 | Helped |
| add LCM (ablation) | CAT-R · Urban100 x2 | PSNR: 32.91 -> 32.98 | Helped |
| use axial-Rwin with small side length sl=2 instead of regular-Rwin (ablation) | CAT · Urban100 x2 | PSNR: 32.98 -> 32.97 | No clear effect |
| increase axial-Rwin side length sl from 2 to 4 (ablation) | CAT-A · Urban100 x2 | PSNR: 32.97 -> 33.20 | Helped |
| CAT-A vs SwinIR | image SR · Urban100 x2 | PSNR: +0.45 dB | Helped |
| CAT vs Restormer | real image denoising · SIDD | PSNR: 40.02 -> 40.01 | No clear effect |
| CAT vs SwinIR | JPEG artifact reduction · Classic5 q=10 | PSNR: 30.27 -> 30.26 | No clear effect |
Closest papers by meaning
- Accurate Image Restoration with Attention Retractable Transformer (ICLR 2023 · similarity 0.88)
- Sharing Key Semantics in Transformer Makes Efficient Image Restoration (NeurIPS 2024 · similarity 0.87)
- Efficient and Explicit Modelling of Image Hierarchies for Image Restoration (CVPR 2023 · similarity 0.87)
- Recursive Generalization Transformer for Image Super-Resolution (ICLR 2024 · similarity 0.86)
- Adapt or Perish: Adaptive Sparse Transformer with Attentive Feature Refinement for Image Restoration (CVPR 2024 · similarity 0.85)
- Intra and Inter Parser-Prompted Transformers for Effective Image Restoration (AAAI 2025 · similarity 0.84)
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.