Research map / Image generation / Image restoration
Removing Rain From Single Images via a Deep Detail Network
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 negative-residual mapping instead of direct image-to-clean-image regression (ablation) | Rain removal CNN · synthetic bird example | SSIM: 0.490 -> 0.926 | Helped |
| add ResNet parameter layers to negative-residual mapping (ablation) | Rain removal CNN · synthetic bird example | SSIM: 0.926 -> 0.936 | Helped |
| add the detail layer as the network input (ablation) | Neg-map+ResNet · synthetic bird example | SSIM: 0.936 -> 0.940 | Helped |
| increase network depth from L = 14 to L = 50 while keeping filter numbers a1 = a2 = 16 (ablation) | Deep detail network · 100 synthetic testing images | SSIM: 0.906 -> 0.921 | Helped |
| increase filter numbers from a1 = a2 = 16 to a1 = a2 = 64 at L = 14 (ablation) | Deep detail network · 100 synthetic testing images | SSIM: 0.906 -> 0.915 | Helped |
| increase network depth from 8 to 50 layers (ablation) | Deep detail network · 100 synthetic testing images | SSIM: 0.896 -> 0.921 | Helped |
| increase ResNet depth from 8 to 50 layers (ablation) | ResNet · 100 synthetic testing images | SSIM: 0.896 -> 0.917 | Helped |
| use the deep detail network rather than ResNet at depth 26 (ablation) | Rain removal network · 100 synthetic testing images | SSIM: 0.907 -> 0.916 | Helped |
| use the deep detail network rather than ResNet at depth 50 (ablation) | Rain removal network · 100 synthetic testing images | SSIM: 0.917 -> 0.921 | Helped |
| increase filter count from 32 to 64 at L = 26 (ablation) | Deep detail network · 100 synthetic testing images | SSIM: 0.920 -> 0.920 | No clear effect |
Closest papers by meaning
- A Model-Driven Deep Neural Network for Single Image Rain Removal (CVPR 2020 · similarity 0.93)
- Deep Joint Rain Detection and Removal From a Single Image (CVPR 2017 · similarity 0.93)
- Rain Streak Removal via Dual Graph Convolutional Network (AAAI 2021 · similarity 0.93)
- Singe Image Rain Removal with Unpaired Information: A Differentiable Programming Perspective (AAAI 2019 · similarity 0.92)
- Depth-Attentional Features for Single-Image Rain Removal (CVPR 2019 · similarity 0.92)
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution (AAAI 2025 · similarity 0.90)
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.