Research map / Detection and segmentation / Open-set and few-shot detection
Curriculum Model Adaptation with Synthetic and Real Data for Semantic Foggy Scene Understanding
Dengxin Dai, Christos Sakaridis, Simon Hecker, Luc Van Gool
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 |
|---|---|---|---|
| add fog densification (CMAda3+ vs CMAda3) (ablation) | CMAda3 · Foggy Zurich-test | mIoU: 43.7 -> 46.8 | Improved |
| add fog densification (CMAda3+ vs CMAda3) (ablation) | CMAda3 · Foggy Driving | mIoU: 48.9 -> 49.8 | Improved |
| add fog densification (CMAda3+ vs CMAda3) (ablation) | CMAda3 · Foggy Driving-dense | mIoU: 40.6 -> 43.0 | Improved |
| add third adaptation stage (CMAda3 vs CMAda2) (ablation) | CMAda2 · Foggy Driving-dense | mIoU: 37.3 -> 40.6 | Improved |
| use our CNN fog density estimator instead of FADE (ablation) | CMAda1 real fog · Foggy Zurich-test | mIoU: 37.5 -> 38.9 | Improved |
| use our CNN fog density estimator instead of FADE (ablation) | CMAda2 (SDBF) · Foggy Zurich-test | mIoU: 40.6 -> 42.9 | Improved |
| use semantic-aware SDBF fog instead of SGF (ablation) | CMAda1 · Foggy Zurich-test | mIoU: 35.7 -> 36.3 | Improved |
| use SDBF instead of SGF in two-stage pipeline (ablation) | CMAda2 · Foggy Zurich-test | mIoU: 41.5 -> 42.9 | Improved |
| use SDBF instead of SGF in two-stage pipeline (ablation) | CMAda2 · Foggy Driving-dense | mIoU: 37.0 -> 37.3 | No clear effect |
| add real fog with density ranking to synthetic stage (two-stage vs single-stage) (ablation) | CMAda1 · Foggy Zurich-test | mIoU: 36.3 -> 42.9 | Improved |
| add confidence-threshold pseudo-label selection (ablation) | CMAda self-learning · Foggy Zurich | mIoU | No clear effect |
| add domain-adversarial adaptation (AdSegNet) instead of CMAda | RefineNet · Foggy Zurich-test | mIoU: 34.6 -> 25.0 | Hurt |
Closest papers by meaning
- Learning Generalized Segmentation for Foggy-Scenes by Bi-directional Wavelet Guidance (AAAI 2024 · similarity 0.91)
- Semantic Segmentation on Raindrop Degraded Images Using Two-Stage Dual Teacher-Student Learning (AAAI 2025 · similarity 0.87)
- Parsing All Adverse Scenes: Severity-Aware Semantic Segmentation with Mask-Enhanced Cross-Domain Consistency (AAAI 2024 · similarity 0.87)
- Semantic Segmentation in Multiple Adverse Weather Conditions with Domain Knowledge Retention (AAAI 2024 · similarity 0.86)
- UniMix: Towards Domain Adaptive and Generalizable LiDAR Semantic Segmentation in Adverse Weather (CVPR 2024 · similarity 0.86)
- Synthetic-to-Real Self-supervised Robust Depth Estimation via Learning with Motion and Structure Priors (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.
