Research map / Point clouds / Segmentation and detection
Open-CRB: Toward Open World Active Learning for 3D Object Detection
Zhuoxiao Chen, Yadan Luo, Zixin Wang, Zijian Wang, Zi Helen Huang
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 OLC in the first selection round (ablation) | Random · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP_unk: 43.05 -> 44.44 | Improved |
| add OLC in the first selection round (ablation) | Cider · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP_unk: 36.23 -> 39.89 | Improved |
| add OLC in the first selection round (ablation) | Cider · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP: 54.71 -> 54.63 | No clear effect |
| add OLC in the first selection round (ablation) | GradNorm · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP_unk: 31.43 -> 42.66 | Improved |
| add OLC in the first selection round (ablation) | GradNorm · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP: 52.85 -> 57.80 | Improved |
| add OLC in the first selection round (ablation) | CRB (Open-CRB) · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP_unk: 46.42 -> 49.23 | Improved |
| add OLC in the first selection round (ablation) | CRB (Open-CRB) · KITTI val (OWAL-3D, 1000 queried boxes) | BEV mAP: 66.51 -> 67.71 | Improved |
| add OLC in the first selection round (ablation) | Random · KITTI val (OWAL-3D, 1000 queried boxes) | 3D mAP: 56.09 -> 57.95 | Improved |
| use uncertainty-based selection instead of diversity-based selection | AL baselines on KITTI · KITTI (OWAL-3D) | 3D mAP_unk | Improved |
| use Coreset (diversity) instead of Open-CRB | nuScenes OWAL-3D · nuScenes (OWAL-3D) | mAP_unk: 2.5 times the annotation cost | Mixed |
| OLC first-round selection vs PROB | Open-CRB · KITTI (OWAL-3D) | BEV mAP_H: 52.64 vs 52.27 | No clear effect |
| use CRB selection instead of other AL methods | CWAL-3D PV-RCNN · KITTI val (CWAL-3D, 1% boxes) | 3D mAP: 6.7% | Improved |
Closest papers by meaning
- Training an Open-Vocabulary Monocular 3D Detection Model without 3D Data (NeurIPS 2024 · similarity 0.91)
- STONE: A Submodular Optimization Framework for Active 3D Object Detection (NeurIPS 2024 · similarity 0.91)
- ImOV3D: Learning Open Vocabulary Point Clouds 3D Object Detection from Only 2D Images (NeurIPS 2024 · similarity 0.90)
- Towards 3D Objectness Learning in an Open World (NeurIPS 2025 · similarity 0.90)
- AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds (AAAI 2025 · similarity 0.89)
- iDet3D: Towards Efficient Interactive Object Detection for LiDAR Point Clouds (AAAI 2024 · similarity 0.89)
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.
