Research map / Label-efficient and robust learning / Domain adaptation and generalization
DODA: Data-Oriented Sim-to-Real Domain Adaptation for 3D Semantic Segmentation
Runyu Ding, Jihan Yang, Li Jiang, Xiaojuan Qi
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 occlusion simulation of virtual scan simulation (ablation) | Source Only · 3D-FRONT -> ScanNet (11 classes) | mIoU: 29.60 -> 39.25 | Improved |
| add noise simulation to occlusion simulation (ablation) | Source Only · 3D-FRONT -> ScanNet (11 classes) | mIoU: 39.25 -> 40.52 | Improved |
| add tail-aware cuboid mixing (TACM) on top of VSS self-training (ablation) | Noisy Student + VSS · 3D-FRONT -> ScanNet (11 classes) | mIoU: 48.13 -> 51.42 | Improved |
| add cuboid mixing (ablation) | Noisy Student + VSS · 3D-FRONT -> ScanNet (11 classes) | mIoU: 48.13 -> 50.55 | Improved |
| add tail cuboid over-sampling (ablation) | Noisy Student + VSS + cuboid mixing · 3D-FRONT -> ScanNet (11 classes) | mIoU: 50.55 -> 51.42 | Improved |
| add virtual scan simulation to UDA baselines (ablation) | CBST · 3D-FRONT -> ScanNet (11 classes) | mIoU: 36.30 -> 47.70 | Improved |
| add virtual scan simulation to UDA baselines (ablation) | MCD · 3D-FRONT -> ScanNet (11 classes) | mIoU: 32.37 -> 40.32 | Improved |
| add TACM to CBST in cross-site adaptation (ablation) | CBST · ScanNet -> S3DIS (8 classes) | mIoU: 60.13 -> 65.52 | Improved |
| add TACM to Noisy Student in cross-site adaptation (ablation) | Noisy Student · S3DIS -> ScanNet (8 classes) | mIoU: 44.81 -> 48.47 | Improved |
| use fewer cameras (2 instead of 4) in VSS (ablation) | VSS · 3D-FRONT -> ScanNet (11 classes) | mIoU: 40.52 -> 38.80 | Hurt |
| use perspective viewing mode instead of fixed in VSS (ablation) | VSS · 3D-FRONT -> ScanNet (11 classes) | mIoU: 40.52 -> 39.04 | Hurt |
| replace TACM with CutMix (ablation) | DODA · 3D-FRONT -> ScanNet (11 classes) | mIoU: 51.42 -> 49.19 | Hurt |
Closest papers by meaning
- Geometry-Aware Network for Domain Adaptive Semantic Segmentation (AAAI 2023 · similarity 0.89)
- Cross-Modal Learning for Domain Adaptation in 3D Semantic Segmentation (TPAMI 2022 · similarity 0.89)
- UniDxMD: Towards Unified Representation for Cross-Modal Unsupervised Domain Adaptation in 3D Semantic Segmentation (ICCV 2025 · similarity 0.88)
- Cross-Modal Contrastive Learning for Domain Adaptation in 3D Semantic Segmentation (AAAI 2023 · similarity 0.88)
- Density-guided Translator Boosts Synthetic-to-Real Unsupervised Domain Adaptive Segmentation of 3D Point Clouds (CVPR 2024 · similarity 0.87)
- UniDSeg: Unified Cross-Domain 3D Semantic Segmentation via Visual Foundation Models Prior (NeurIPS 2024 · 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.
