Research map / Detection and segmentation / Open-set and few-shot detection
Rain Rendering for Evaluating and Improving Robustness to Bad Weather
Maxime Tremblay, Shirsendu Sukanta Halder, Raoul de Charette, Jean‐François Lalonde
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 |
|---|---|---|---|
| fine-tune YOLOv2 on PBR rain with curriculum | YOLOv2 · real nuScenes-rain | mAP (%): 16.30 -> 19.68 (+20.7%) | Improved |
| fine-tune YOLOv2 on GAN rain | YOLOv2 · real nuScenes-rain | mAP (%): 16.30 -> 18.07 (+10.9%) | Improved |
| fine-tune YOLOv2 on GAN+PBR rain | YOLOv2 · real nuScenes-rain | mAP (%): 16.30 -> 19.73 (+21.0%) | Improved |
| fine-tune PSPNet on PBR rain | PSPNet · real nuScenes-rain | AP (%): 18.7 -> 25.6 (+36.9%) | Improved |
| fine-tune Monodepth2 on PBR rain | Monodepth2 · real nuScenes-rain | Sq. rel. err. (%): 3.53 -> 3.54 (0.0%) | No clear effect |
| fine-tune Monodepth2 on GAN+PBR rain | Monodepth2 · real nuScenes-rain | Sq. rel. err. (%): 3.53 -> 3.29 (+8.2%) | Improved |
| fine-tune PSPNet on PBR rain, effect on clear images | PSPNet · real nuScenes-clear | AP (%): 40.8 -> 39.0 | Hurt |
| fine-tune YOLOv2 with rain curriculum (PBR), robustness at 200 mm/hr | YOLOv2 · nuScenes-augment PBR | mAP (%): 20.4 -> 31.0 | Improved |
| fine-tune PSPNet with rain curriculum, at 100 mm/hr | PSPNet · Cityscapes PBR 100 mm/hr | AP (%): 52.0 -> 54.0 | Improved |
| fine-tune YOLOv2 on hybrid GAN+PBR at 100 mm/hr | YOLOv2 · nuScenes-augment GAN+PBR 100 mm/hr | mAP (%): 7.5 -> 21.7 | Improved |
| fine-tune Monodepth2 on hybrid GAN+PBR at 100 mm/hr | Monodepth2 · nuScenes-augment GAN+PBR 100 mm/hr | Sq. rel. err. (%): 8.9 -> 5.7 | Improved |
| de-rain input with DualResNet instead of rain-aware finetuning, depth | Monodepth2 · real nuScenes-rain | Sq. rel. err. (%): 3.29 -> 3.09 | Improved |
Closest papers by meaning
- Semantic Segmentation on Raindrop Degraded Images Using Two-Stage Dual Teacher-Student Learning (AAAI 2025 · similarity 0.88)
- Towards Universal Rainy Image Restoration: Benchmark and Baseline (AAAI 2025 · similarity 0.88)
- Harnessing Joint Rain-/Detail-aware Representations to Eliminate Intricate Rains (ICLR 2024 · similarity 0.88)
- Sunshine to Rainstorm: Cross-Weather Knowledge Distillation for Robust 3D Object Detection (AAAI 2024 · similarity 0.88)
- ERF: A Benchmark Dataset for Robust Semantic Segmentation Under Extreme Rainfall Conditions (AAAI 2025 · similarity 0.87)
- RainyGS: Efficient Rain Synthesis with Physically-Based Gaussian Splatting (CVPR 2025 · similarity 0.87)
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.
