Research map / Detection and segmentation / 3D detection and driving
Harnessing Uncertainty-aware Bounding Boxes for Unsupervised 3D Object Detection
Ruiyang Zhang, Hu Zhang, Zhedong Zheng
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 UA3D uncertainty estimation and regularization to self-training (vs OYSTER) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 12.7 -> 19.6 | Improved |
| add UA3D (vs MODEST) | UA3D · Lyft | AP_BEV@0.25 (0-80m): 57.3 -> 61.4 | Improved |
| add UA3D on long range (50-80m) | UA3D · nuScenes | AP_BEV@0.25 (50-80m): 1.7 -> 4.3 | Improved |
| add UA3D on long range (50-80m) | UA3D · Lyft | AP_BEV@0.25 (50-80m): 27.0 -> 36.8 | Improved |
| use learnable uncertainty instead of distance-based rule uncertainty (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 14.8 -> 19.6 | Improved |
| use learnable uncertainty instead of num-points-based rule uncertainty (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 13.7 -> 19.6 | Improved |
| use learnable uncertainty instead of volume-based rule uncertainty (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 12.3 -> 19.6 | Improved |
| use coordinate-level instead of box-level uncertainty (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 17.2 -> 19.6 | Improved |
| use coordinate-level instead of point-cloud-level uncertainty (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 12.1 -> 19.6 | Improved |
| set auxiliary detector channel coefficient gamma=0.5 instead of 0.25 (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 16.9 -> 19.6 | Improved |
| set gamma=0.5 instead of 1 (same-capacity auxiliary) (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 14.7 -> 19.6 | Improved |
| set lambda=1e-5 instead of 1e-4 (ablation) | UA3D · nuScenes | AP_BEV@0.25 (0-80m): 15.2 -> 19.6 | Improved |
Closest papers by meaning
- Not Every Side Is Equal: Localization Uncertainty Estimation for Semi-Supervised 3D Object Detection (ICCV 2023 · similarity 0.89)
- Uni3D: A Unified Baseline for Multi-Dataset 3D Object Detection (CVPR 2023 · similarity 0.89)
- NBA3D: Neighbor-Based Confidence Adjustment for 3D Rare Object Detection Using LiDAR (AAAI 2025 · similarity 0.87)
- UNION: Unsupervised 3D Object Detection using Object Appearance-based Pseudo-Classes (NeurIPS 2024 · similarity 0.86)
- STONE: A Submodular Optimization Framework for Active 3D Object Detection (NeurIPS 2024 · similarity 0.86)
- SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection (CVPR 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.
