Research map / Detection and segmentation / 3D detection and driving
Towards Stable 3D Object Detection
Jiabao Wang, Qiang Meng, Guochao Liu, Liujiang Yan, Ke Wang, Ming-Ming Cheng, Qibin Hou
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 |
|---|---|---|---|
| use a more accurate detector (higher mAPH) | 3D detectors on Waymo · Waymo Open Dataset val, Vehicle | SI: TransFusion mAPH 79.00, SI 82.32 | No clear effect |
| fuse camera images (TransFusion) with LiDAR | 3D detector · Waymo Open Dataset val, Vehicle | SI: SI 82.32 vs DSVT 84.90 despite higher mAPH 79.00 | Hurt |
| increase input frames from 1 to 4 (ablation) | CenterPoint · Waymo Open Dataset val, Vehicle | SI: 80.52 -> 81.74 | Improved |
| increase input frames from 1 to 4 (ablation) | PV R-CNN · Waymo Open Dataset val, Vehicle | SI: 85.17 -> 87.50 | Improved |
| fine-tune CenterPoint without consistency loss (ablation) | CenterPoint · Waymo Open Dataset val, Vehicle | SI: 80.52 -> 80.93 | No clear effect |
| add prediction consistency learning (n=0) (ablation) | CenterPoint · Waymo Open Dataset val, Vehicle | SI: 80.52 -> 85.42 | Improved |
| increase max frame interval n from 0 to 16 in PCL (ablation) | CenterPoint + PCL · Waymo Open Dataset val, Vehicle | SI: 85.42 -> 86.00 | Improved |
| increase max frame interval n from 0 to 16 in PCL (ablation) | CenterPoint + PCL · Waymo Open Dataset val, Vehicle | mAPH: 75.57 -> 74.54 | Hurt |
| add PCL (n=0) | CenterPoint · Waymo Open Dataset val, Cyclist | mAPH: 71.04 -> 70.86 | Hurt |
| add confidence consistency loss alone (ablation) | PCL · Waymo Open Dataset val, Vehicle | SI: 80.93 -> 84.04 | Improved |
| add heading consistency loss alone (ablation) | PCL · Waymo Open Dataset val, Vehicle | SI: 80.93 -> 80.89 | No clear effect |
| combine all four loss parts (C, L, E, H) (ablation) | PCL · Waymo Open Dataset val, Vehicle | SI: 84.04 -> 85.42 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- TransFusion — TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection With Transformers (CVPR 2022) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Weakly Supervised Monocular 3D Object Detection by Spatial-Temporal View Consistency (TPAMI 2024 · similarity 0.88)
- Benchmarking Robustness of 3D Object Detection to Common Corruptions (CVPR 2023 · similarity 0.88)
- Towards Building Self-Aware Object Detectors via Reliable Uncertainty Quantification and Calibration (CVPR 2023 · similarity 0.87)
- Bounding Box Stability against Feature Dropout Reflects Detector Generalization across Environments (ICLR 2024 · similarity 0.86)
- OpenAD: Open-World Autonomous Driving Benchmark for 3D Object Detection (NeurIPS 2025 · similarity 0.85)
- NBA3D: Neighbor-Based Confidence Adjustment for 3D Rare Object Detection Using LiDAR (AAAI 2025 · similarity 0.85)
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.
