Research map / Image generation / Image restoration
Deep Stacked Hierarchical Multi-Patch Network for Image 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 |
|---|---|---|---|
| Use a fine-to-coarse multi-patch hierarchy rather than the matched multi-scale baseline | DMPHN · GoPro test | PSNR: 28.97 -> 30.21 | Helped |
| Stack two DMPHN units (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 30.71 | Helped |
| Stack four DMPHN units (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 31.20 | Helped |
| Use the reversed top-bottom-top VMPHN stacking direction (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 30.90 | Helped |
| Stack two VMPHN units (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 31.50 | Helped |
| Add the 16-patch finest level (ablation) | DMPHN(1-2-4-8) · GoPro test | PSNR: 30.25 -> 29.87 | Hurt |
| Use hierarchy (1-4-16) instead of (1-2-4) (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 29.15 | Hurt |
| Use hierarchy (1-1-1) instead of (1-2-4) (ablation) | DMPHN · GoPro test | PSNR: 30.21 -> 28.11 | Hurt |
| Share encoder-decoder weights across hierarchy levels (ablation) | DMPHN(1-2) · GoPro test | PSNR: 29.77 -> 29.22 | Hurt |
| Share encoder-decoder weights across hierarchy levels (ablation) | DMPHN(1-2-4) · GoPro test | PSNR: 30.21 -> 29.56 | Hurt |
| Share encoder-decoder weights across hierarchy levels (ablation) | DMPHN(1-2-4-8) · GoPro test | PSNR: 30.25 -> 30.04 | Hurt |
| Apply MSE loss at multiple hierarchy levels instead of only the final output (ablation) | DMPHN · GoPro | deblurring performance | No clear effect |
Closest papers by meaning
- Spatially-Attentive Patch-Hierarchical Network for Adaptive Motion Deblurring (CVPR 2020 · similarity 0.89)
- Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring (CVPR 2024 · similarity 0.88)
- GS-Blur: A 3D Scene-Based Dataset for Realistic Image Deblurring (NeurIPS 2024 · similarity 0.86)
- Parameterized Blur Kernel Prior Learning for Local Motion Deblurring (CVPR 2025 · similarity 0.86)
- BlurDM: A Blur Diffusion Model for Image Deblurring (NeurIPS 2025 · similarity 0.85)
- Exploiting Deblurring Networks for Radiance Fields (CVPR 2025 · 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.