3D Object Detection on Waymo Open Dataset Pedestrians (val)
82.88L1 3D AP (IoU=0.5)3D Auto Labeling
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| 3D Auto Labelingframes=all2021.03 | 82.88 | 63.69 | 86.32 | 75.6 | |
| MVF++frames=5, TTA=true2021.03 | 81.83 | 60.56 | 85.9 | 73 | |
| MVF++frames=12021.03 | 78.01 | 56.02 | 83.31 | 68.04 | |
| PillarNetframes=12021.03 | 72.51 | — | 78.53 | — | |
| StarNetframes=12021.03 | 66.8 | — | — | — | |
| PV-RCNNframes=12021.03 | 65.34 | 45.12 | 70.35 | 56.63 | |
| Multi-view fusion (MVF)frames=12021.03 | 65.33 | — | 74.38 | — | |
| PointPillarframes=12021.03 | 60.11 | 40.35 | 65.42 | 51.71 |