3D Object Detection on KITTI cross-dataset from PandaSet (val)
76.7AP_BEVUltraLiDAR (Ours) + Two-stage PIXOR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| UltraLiDAR (Ours) + Two-stage PIXORdetector=Two-stage PIXOR, input_data=Sim / 512, sparse_to_dense=UltraLiDAR2023.11 | 76.7 | 46.3 | |
| Two-stage PIXOR + ContCompdetector=Two-stage PIXOR, input_data=Sim / 512, sparse_to_dense=ContComp2023.11 | 74.9 | 41.5 | |
| UltraLiDAR (Ours) + PointPillardetector=PointPillar, input_data=Sim / 512, sparse_to_dense=UltraLiDAR2023.11 | 73 | 40.9 | |
| Two-stage PIXORdetector=Two-stage PIXOR, input_data=Real / 64, sparse_to_dense=None2023.11 | 71.7 | 32.8 | |
| PointPillar + ContCompdetector=PointPillar, input_data=Sim / 512, sparse_to_dense=ContComp2023.11 | 67.7 | 36.9 | |
| Two-stage PIXORdetector=Two-stage PIXOR, input_data=Sim / 512, sparse_to_dense=None2023.11 | 66.9 | 33.2 | |
| PointPillardetector=PointPillar, input_data=Real / 64, sparse_to_dense=None2023.11 | 60.9 | 28.1 | |
| PointPillardetector=PointPillar, input_data=Sim / 512, sparse_to_dense=None2023.11 | 58.5 | 28 |