Research map / Label-efficient and robust learning / OOD and anomaly detection
ProOOD: Prototype-Guided Out-of-Distribution 3D Occupancy Prediction
Yuheng Zhang, Mengfei Duan, Kunyu Peng, Yuhang Wang, Di Wen, Danda Pani Paudel, Luc Van Gool, Kailun Yang
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 prototype-based contrastive loss (PBCL) (ablation) | SGN baseline · SemanticKITTI val | Tail mIoU: 4.07 -> 4.44 | Improved |
| add Prototype-Guided Semantic Imputation (PGSI) (ablation) | SGN + PBCL · SemanticKITTI val | IoU: 43.69 -> 44.09 | Improved |
| add Prototype-Guided Semantic Imputation (PGSI) (ablation) | SGN + PBCL · SemanticKITTI val | Tail mIoU: 4.44 -> 4.58 | Improved |
| add Prototype-Guided Tail Mining (PGTM) (ablation) | SGN + PBCL + PGSI · SemanticKITTI val | mIoU: 14.76 -> 14.97 | Improved |
| add Prototype-Guided Tail Mining (PGTM) (ablation) | SGN + PBCL + PGSI · SemanticKITTI val | Tail mIoU: 4.58 -> 4.89 | Improved |
| add Prototype-Guided Tail Mining (PGTM) (ablation) | SGN + PBCL + PGSI · SemanticKITTI val | IoU: 44.09 -> 44.02 | No clear effect |
| use EchoOOD score instead of ASS (ablation) | ProOOD (+SGN) · VAA-KITTI | AuPRC_r (1.0m): 20.96 -> 43.55 | Improved |
| use EchoOOD score instead of ASS (ablation) | ProOOD (+SGN) · VAA-KITTI | AuROC: 64.20 -> 64.31 | No clear effect |
| use EchoOOD score instead of Energy (ablation) | ProOOD (+SGN) · VAA-KITTI | AuROC: 66.69 -> 64.31 | Hurt |
| set EMA momentum to 0.05 instead of 0.01 (ablation) | ProOOD (+SGN) · SemanticKITTI val | Tail mIoU: 4.32 -> 4.89 | Improved |
| raise EMA momentum from 0.05 to 0.2 (ablation) | ProOOD (+SGN) · SemanticKITTI val | Tail mIoU: 4.89 -> 4.36 | Hurt |
| reduce margin delta from 0.10 to 0.05 (ablation) | ProOOD (+SGN) · SemanticKITTI val | Tail mIoU: 4.89 -> 4.25 | Hurt |
Closest papers by meaning
- Achieving Speed-Accuracy Balance in Vision-based 3D Occupancy Prediction via Geometric-Semantic Disentanglement (AAAI 2025 · similarity 0.89)
- OPUS: Occupancy Prediction Using a Sparse Set (NeurIPS 2024 · similarity 0.89)
- OctOcc: High-Resolution 3D Occupancy Prediction with Octree (AAAI 2024 · similarity 0.89)
- EvOcc: Accurate Semantic Occupancy for Automated Driving Using Evidence Theory (CVPR 2025 · similarity 0.88)
- CSV-Occ: Fusing Multi-frame Alignment for Occupancy Prediction with Temporal Cross State Space Model and Central Voting Mechanism (ICML 2025 · similarity 0.88)
- Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion in Connected Automated Vehicles (CVPR 2024 · 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.
