Research map / Detection and segmentation / Open-set and few-shot detection
DANNet: A One-Stage Domain Adaptation Network for Unsupervised Nighttime Semantic Segmentation
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 DANNet with PSPNet instead of the strongest reported baseline, MGCDA | Nighttime semantic segmentation · Dark Zurich-test | mIoU: 42.5 -> 45.2 | Helped |
| use DANNet with PSPNet instead of MGCDA | Nighttime semantic segmentation · Nighttime Driving test | mIoU: 49.4 -> 47.70 | Hurt |
| remove Dark Zurich-D data (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 22.78 | Hurt |
| remove the relighting network and L_light (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 34.14 | Hurt |
| remove L_light (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 35.05 | Hurt |
| remove L_static (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 20.48 | Hurt |
| replace the specially designed static loss with cross-entropy loss (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 33.61 | Hurt |
| replace the specially designed static loss with focal loss (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 36.49 | No clear effect |
| remove re-weighting on pseudo labels (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 32.71 | Hurt |
| remove re-weighting on predictions (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 32.22 | Hurt |
| remove the pretrained segmentation model (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU: 36.76 -> 30.74 | Hurt |
| set the re-weighting std to 0.16 for testing (ablation) | DANNet (PSPNet) · Dark Zurich-val | mIoU | Helped |
Closest papers by meaning
- The Parables of the Mustard Seed and the Yeast: Extremely Low-Budget, High-Performance Nighttime Semantic Segmentation (AAAI 2025 · similarity 0.91)
- Cross-Domain Correlation Distillation for Unsupervised Domain Adaptation in Nighttime Semantic Segmentation (CVPR 2022 · similarity 0.91)
- NightAdapter: Learning a Frequency Adapter for Generalizable Night-time Scene Segmentation (CVPR 2025 · similarity 0.86)
- Exploring the Common Appearance-Boundary Adaptation for Nighttime Optical Flow (ICLR 2024 · similarity 0.85)
- Day-Night Cross-domain Vehicle Re-identification (CVPR 2024 · similarity 0.84)
- NightRain: Nighttime Video Deraining via Adaptive-Rain-Removal and Adaptive-Correction (AAAI 2024 · similarity 0.84)
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.