BEV Object Detection on Waymo Open Dataset Vehicles (val)
86.1AP L1Pillar-based
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Pillar-basedpipeline=two-stage2021.06 | 86.1 | — | |
| RCNpipeline=two-stage2021.06 | 83.4 | — | |
| PV-RCNNpipeline=two-stage2021.06 | 83 | 67.6 | |
| PPBA-PointPillars2021.06 | 81.4 | 72.2 | |
| PPC + EdgeConvkernel=EdgeConv2021.06 | 80.8 | 71.8 | |
| PPC + PointNetkernel=PointNet2021.06 | 80.5 | 71.6 | |
| MultiView2021.06 | 80.4 | — | |
| PPC + Conv2Dkernel=Conv2D2021.06 | 78.1 | 68.9 | |
| PointPillarsreimplemented=true2021.06 | 77.2 | 67.6 | |
| PPC + RQ-Conv2Dkernel=RQ-Conv2D2021.06 | 76.2 | 67.2 | |
| LaserNetreimplemented=true2021.06 | 73.1 | 63.9 | |
| StarNet2021.06 | 67.7 | 60 |