Research map / Label-efficient and robust learning / Domain adaptation and generalization
Map-Guided Curriculum Domain Adaptation and Uncertainty-Aware Evaluation for Semantic Nighttime Image Segmentation
Christos Sakaridis, Dengxin Dai, 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 twilight data to nighttime adaptation (no curriculum) (ablation) | single-stage adaptation · Dark Zurich-test | mIoU: 33.6 -> 34.0 | No clear effect |
| add two-stage curriculum with twilight (ablation) | adaptation w/ twilight · Dark Zurich-test | mIoU: 34.0 -> 38.2 | Improved |
| add guided refinement with cross bilateral filtering (ablation) | MGCDA w/o guided refinement · Dark Zurich-test | mIoU: 38.2 -> 40.3 | Improved |
| use depth-based warping instead of cross bilateral filtering (ablation) | MGCDA · Dark Zurich-test | mIoU: 40.3 -> 42.5 | Improved |
| train CycleGAN on full 360x720 images instead of 256x256 crops (ablation) | synthetic adaptation · Dark Zurich-test | mIoU: 37.1 -> 40.2 | Improved |
| add map guidance at test time (ablation) | MGCDA · Dark Zurich-test | mIoU: 42.5 -> 44.1 | Improved |
| add map guidance at test time (ablation) | DMAda · Dark Zurich-test | mIoU: 32.1 -> 34.8 | Improved |
| preprocess nighttime images with CycleGAN before daytime model (ablation) | RefineNet · Dark Zurich-test | mIoU: 28.5 -> 7.2 | Hurt |
| preprocess with CLAHE (ablation) | RefineNet · Dark Zurich-test | mIoU: 28.5 -> 23.0 | Hurt |
| preprocess with ZeroDCE (ablation) | RefineNet · Dark Zurich-test | mIoU: 28.5 -> 17.9 | Hurt |
| enhance synthetic nighttime images with ZeroDCE during adaptation (ablation) | RefineNet adaptation · Dark Zurich-test | mIoU: 40.2 -> 38.4 | Hurt |
| use MGCDA instead of GCMA | adaptation pipeline · Nighttime Driving | mIoU: 45.6 -> 49.4 | Improved |
Closest papers by meaning
- Guided Curriculum Model Adaptation and Uncertainty-Aware Evaluation for Semantic Nighttime Image Segmentation (ICCV 2019 · similarity 0.98)
- The Parables of the Mustard Seed and the Yeast: Extremely Low-Budget, High-Performance Nighttime Semantic Segmentation (AAAI 2025 · similarity 0.93)
- NightAdapter: Learning a Frequency Adapter for Generalizable Night-time Scene Segmentation (CVPR 2025 · similarity 0.88)
- NightCC: Nighttime Color Constancy via Adaptive Channel Masking (CVPR 2024 · similarity 0.87)
- Exploring the Common Appearance-Boundary Adaptation for Nighttime Optical Flow (ICLR 2024 · similarity 0.86)
- See through the Dark: Learning Illumination-affined Representations for Nighttime Occupancy Prediction (NeurIPS 2025 · similarity 0.86)
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.
