Research map / Architectures and efficiency / CNN design
Spatially-Attentive Patch-Hierarchical Network for Adaptive Motion Deblurring
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 |
|---|---|---|---|
| Add adaptive kernel filtering to the network without attention (ablation) | Proposed network · GoPro test set | PSNR: 30.25 -> 30.81 | Helped |
| Add self-attention (SA) without adaptive filtering (ablation) | Proposed network · GoPro test set | PSNR: 30.25 -> 30.76 | Helped |
| Add cross-attention (CA) to the SA network (ablation) | Proposed network · GoPro test set | PSNR: 30.76 -> 30.93 | Helped |
| Add cross-level attention (CLA) to the SA network (ablation) | Proposed network · GoPro test set | PSNR: 30.76 -> 31.12 | Helped |
| Add adaptive kernel filtering to the SA+CA network (ablation) | Proposed network · GoPro test set | PSNR: 30.93 -> 31.44 | Helped |
| Add CLA to the SA+CA+adaptive-kernel network (ablation) | Proposed network · GoPro test set | PSNR: 31.44 -> 31.85 | Helped |
| Add learnable local-filter offsets (ablation) | Proposed network · GoPro test set | PSNR: 31.85 -> 32.02 | Helped |
| Replace the proposed efficient attention with the compared standard attention (ablation) | Net3 · GoPro test set | PSNR: 30.76 -> 30.52 | Hurt |
| Increase encoder-decoder residual blocks beyond 3 (ablation) | Proposed network · GoPro test set | PSNR | No clear effect |
| Add a fourth hierarchy level (ablation) | Proposed network · Deblurring benchmarks | PSNR | No clear effect |
Closest papers by meaning
- Parameterized Blur Kernel Prior Learning for Local Motion Deblurring (CVPR 2025 · similarity 0.90)
- Motion-adaptive Transformer for Event-based Image Deblurring (AAAI 2025 · similarity 0.89)
- Adaptive Window Pruning for Efficient Local Motion Deblurring (ICLR 2024 · similarity 0.89)
- Real-World Efficient Blind Motion Deblurring via Blur Pixel Discretization (CVPR 2024 · similarity 0.88)
- Motion-adaptive Separable Collaborative Filters for Blind Motion Deblurring (CVPR 2024 · similarity 0.88)
- DeblurDiff: Real-Word Image Deblurring with Generative Diffusion Models (NeurIPS 2025 · similarity 0.87)
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.