Research map / Image generation / Image restoration
From Rain Generation to Rain Removal
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 |
|---|---|---|---|
| Augment PReNet training with VRGNet-generated rain pairs | PReNet on Rain100L · Rain100L | PSNR: 37.42 -> 37.84 | Helped |
| Augment PReNet training with VRGNet-generated rain pairs | PReNet on Rain100H · Rain100H | PSNR: 30.11 -> 30.48 | Helped |
| Augment PReNet training with VRGNet-generated rain pairs | PReNet on Rain1400 · Rain1400 | PSNR: 32.21 -> 32.51 | Helped |
| Augment PReNet training with VRGNet-generated rain pairs | PReNet on SPA-Data · SPA-Data | PSNR: 40.16 -> 40.27 | Helped |
| Augment PReNet training with VRGNet-generated rain pairs | PReNet trained on Rain100L · SPA-Data test (cross-dataset generalization) | PSNR: 34.91 -> 35.13 | Helped |
| Use generated instead of additional real training pairs, with a total of 3K pairs | PReNet on SPA-Data · SPA-Data test | PSNR: 39.96 -> 40.25 | Helped |
| Remove BNet and use the estimated clean image as the background (ablation) | VRGNet augmentation with PReNet · SPA-Data test, 1K+0.5K samples | PSNR: 39.71 -> 39.41 | Hurt |
| Remove BNet and use the estimated clean image as the background (ablation) | VRGNet augmentation with PReNet · SPA-Data test, 1K+1K samples | PSNR: 39.83 -> 39.39 | Hurt |
| Remove BNet and use the estimated clean image as the background (ablation) | VRGNet augmentation with PReNet · SPA-Data test, 1K+2K samples | PSNR: 40.25 -> 39.35 | Hurt |
| Jointly train VRGNet rather than training BNet alone (ablation) | PReNet on Rain100L · Rain100L | PSNR: 37.42 -> 36.94 | Hurt |
Closest papers by meaning
- EfficientDeRain: Learning Pixel-wise Dilation Filtering for High-Efficiency Single-Image Deraining (AAAI 2021 · similarity 0.91)
- Semi-Supervised Transfer Learning for Image Rain Removal (CVPR 2019 · similarity 0.91)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.90)
- Bidirectional Multi-Scale Implicit Neural Representations for Image Deraining (CVPR 2024 · similarity 0.89)
- Boosting Image De-Raining via Central-Surrounding Synergistic Convolution (AAAI 2025 · similarity 0.88)
- Towards Universal Rainy Image Restoration: Benchmark and Baseline (AAAI 2025 · similarity 0.88)
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.