Research map / Image generation / Image restoration
Density-Aware Single Image De-Raining Using a Multi-Stream Dense Network
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 |
|---|---|---|---|
| use density-guided DID-MDN instead of DDN | DID-MDN · Test1 | SSIM: 0.8978 -> 0.9087 | Helped |
| use density-guided DID-MDN instead of DDN | DID-MDN · Test1 | PSNR (dB): 27.33 -> 27.95 | Helped |
| use density-guided DID-MDN instead of DDN | DID-MDN · Test2 | SSIM: 0.8851 -> 0.9092 | Helped |
| use density-guided DID-MDN instead of DDN | DID-MDN · Test2 | PSNR (dB): 25.63 -> 26.0745 | Helped |
| add estimated rain-density label fusion (ablation) | Multi-stream densely connected network · Test1 | PSNR (dB): 27.56 -> 27.95 | Helped |
| add estimated rain-density label fusion (ablation) | Multi-stream densely connected network · Test1 | SSIM: 0.9028 -> 0.9087 | Helped |
| use a multi-stream dense network instead of one Dense2 stream (ablation) | De-raining network without label fusion · Test1 | PSNR (dB): 26.05 -> 27.56 | Helped |
| use a multi-stream dense network instead of one Dense2 stream (ablation) | De-raining network without label fusion · Test1 | SSIM: 0.8893 -> 0.9028 | Helped |
| use the proposed multi-stream dense structure instead of Yang's multi-scale structure (ablation) | Multi-stream network without label fusion · Test1 | PSNR (dB): 26.75 -> 27.56 | Helped |
| use the proposed multi-stream dense structure instead of Yang's multi-scale structure (ablation) | Multi-stream network without label fusion · Test1 | SSIM: 0.8901 -> 0.9028 | Helped |
| use a residual-aware classifier instead of VGG-16 (ablation) | Rain-density classifier · Test1 rain-density estimation | Accuracy: 73.32 % -> 85.15 % | Helped |
| use DID-MDN instead of DDN | 512X512-image de-raining · 1000-image average | Running time (seconds): 0.3s -> 0.2s | Helped |
Closest papers by meaning
- Deep Joint Rain Detection and Removal From a Single Image (CVPR 2017 · similarity 0.91)
- Removing Rain From Single Images via a Deep Detail Network (CVPR 2017 · similarity 0.91)
- Singe Image Rain Removal with Unpaired Information: A Differentiable Programming Perspective (AAAI 2019 · similarity 0.90)
- Removing Raindrops and Rain Streaks in One Go (CVPR 2021 · similarity 0.90)
- Uncertainty Guided Multi-Scale Residual Learning-Using a Cycle Spinning CNN for Single Image De-Raining (CVPR 2019 · similarity 0.90)
- 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.