Research map / Architectures and efficiency / Neural architecture search
Pyramid Architecture Search for Real-Time Image Deblurring
Xiaobin Hu, Wenqi Ren, Kaicheng Yu, Kaihao Zhang, Xiaochun Cao, Wei Liu, Bjoern Menze
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 |
|---|---|---|---|
| use the PyNAS hierarchical multi-scale search instead of random sampled models (ablation) | multi-scale deblurring network · GoPro test (search strategy comparison) | PSNR: 29.11 -> 30.62 | Improved |
| use PyNAS search instead of NAS-DIP | deblurring network · GoPro test (search strategy comparison) | PSNR: 29.01 -> 30.62 | Improved |
| use PyNAS search instead of CLEARER | deblurring network · GoPro test (search strategy comparison) | PSNR: 28.96 -> 30.62 | Improved |
| search individual cells per scale (PyNASd) instead of shared identical cells (PyNASs) (ablation) | PyNAS · GoPro test | PSNR: 30.51 -> 30.62 | Improved |
| search individual cells per scale (PyNASd) instead of shared cells (PyNASs) (ablation) | PyNAS · GoPro test | runtime (ms): 26 -> 17 | Improved |
| search the patch scheme (1-3-9) instead of fixed manual 1-2-4-8 pyramid (ablation) | PyNAS · GoPro test | PSNR: 30.51 -> 30.62 | Improved |
| replace manual multi-scale design with NAS-found architecture | Zhang et al. DMPHN · GoPro test | PSNR: 30.25 -> 30.62 | Improved |
| replace manual multi-scale design with NAS-found architecture | Zhang et al. DMPHN · VideoDeblurring test | PSNR: 30.76 -> 31.01 | Improved |
| replace manual multi-scale design with NAS-found architecture | Yuan et al. real-time method · GoPro test | PSNR: 29.57 -> 30.62 | Improved |
Closest papers by meaning
- Memory-Efficient Hierarchical Neural Architecture Search for Image Restoration (IJCV 2021 · similarity 0.87)
- Deep Stacked Hierarchical Multi-Patch Network for Image Deblurring (CVPR 2019 · similarity 0.87)
- Scale-Recurrent Network for Deep Image Deblurring (CVPR 2018 · similarity 0.86)
- Densely Connected Pyramid Dehazing Network (CVPR 2018 · similarity 0.86)
- Deep Image Deblurring: A Survey (IJCV 2022 · similarity 0.85)
- Residual Diffusion Deblurring Model for Single Image Defocus Deblurring (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.
