Bird's Eye View Detection on KITTI Car class official (test)
95.94AP (Easy)PointViT V1
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PointViT V1Method Architecture=One-stage, Latency (ms)=36, Evaluation Hardware=1x RTX 4090 GPU, Batch Size=1, Depth Completion Fusion Strategy=Early Fusion + PE-Net2026.02 | 95.94 | 89.95 | 82.4 | |
| PVT-SSDMethod Architecture=One-stage, Latency (ms)=32, Evaluation Hardware=1x RTX 4090 GPU, Batch Size=12026.02 | 95.23 | 91.63 | 86.43 | |
| SA-SSDMethod Architecture=One-stage, Latency (ms)=402026.02 | 95.03 | 91.03 | 85.96 | |
| PV-RCNNMethod Architecture=Two-stage, Latency (ms)=982026.02 | 94.98 | 90.65 | 86.14 | |
| PC-RGNNModality=LiDAR only2020.12 | 94.91 | 89.62 | 86.57 | |
| STDModality=LiDAR only2020.12 | 94.74 | 89.19 | 86.42 | |
| STDMethod Architecture=Two-stage, Latency (ms)=802026.02 | 94.74 | 89.19 | 86.42 | |
| PDVMethod Architecture=Two-stage, Latency (ms)=1352026.02 | 94.56 | 90.48 | 86.23 | |
| EQ-PVRCNNMethod Architecture=Two-stage, Latency (ms)=2002026.02 | 94.55 | 90.09 | 86.42 | |
| M3DETRMethod Architecture=Two-stage2026.02 | 94.41 | 90.37 | 85.98 | |
| PVGNetMethod Architecture=One-stage2026.02 | 94.36 | 91.26 | 86.63 | |
| SPGMethod Architecture=Two-stage, Latency (ms)=1562026.02 | 94.33 | 88.7 | 85.98 | |
| ContFuseModality=RGB+LiDAR2020.12 | 94.07 | 85.35 | 75.88 | |
| VoTr-TSDMethod Architecture=Two-stage, Latency (ms)=2162026.02 | 94.03 | 90.34 | 86.14 | |
| MMFModality=RGB+LiDAR2020.12 | 93.67 | 88.21 | 81.99 | |
| PointViT V2Method Architecture=One-stage, Latency (ms)=36, Evaluation Hardware=1x RTX 4090 GPU, Batch Size=1, Depth Completion Fusion Strategy=Early Fusion + BP-Net2026.02 | 93.59 | 89.67 | 82.12 | |
| PointViT P1Method Architecture=One-stage, Latency (ms)=50, Evaluation Hardware=1x RTX 4090 GPU, Batch Size=1, Depth Completion Fusion Strategy=Early Fusion + BP-Net, LiDAR Input Type=colored raw LiDAR points2026.02 | 93.44 | 89.57 | 82.08 | |
| IA-SSDMethod Architecture=One-stage, Latency (ms)=442026.02 | 93.14 | 89.48 | 84.42 | |
| Point-GNNModality=LiDAR only2020.12 | 93.11 | 89.17 | 83.9 | |
| Point-GNNMethod Architecture=One-stage, Latency (ms)=6432026.02 | 93.11 | 89.17 | 83.19 | |
| SASAMethod Architecture=One-stage, Latency (ms)=632026.02 | 92.87 | 89.51 | 86.35 | |
| 3DSSDModality=LiDAR only2020.12 | 92.66 | 89.02 | 85.86 | |
| 3DSSDMethod Architecture=One-stage, Latency (ms)=382026.02 | 92.66 | 89.02 | 85.86 | |
| CT3DMethod Architecture=Two-stage, Latency (ms)=702026.02 | 92.36 | 88.83 | 84.07 | |
| Pyramid-PVMethod Architecture=Two-stage, Latency (ms)=1272026.02 | 92.19 | 88.84 | 86.21 | |
| PointRCNNModality=LiDAR only2020.12 | 92.13 | 87.39 | 82.72 | |
| PointRCNNMethod Architecture=Two-stage, Latency (ms)=1002026.02 | 92.13 | 87.39 | 82.72 | |
| SVGA-NetMethod Architecture=One-stage, Latency (ms)=622026.02 | 92.07 | 89.08 | 85.59 | |
| PV-RCNN++Method Architecture=Two-stage, Latency (ms)=602026.02 | 92.06 | 88.74 | 85.97 | |
| Part-A2Modality=LiDAR only2020.12 | 91.7 | 87.79 | 84.61 | |
| TANetMethod Architecture=One-stage, Latency (ms)=352026.02 | 91.58 | 86.54 | 81.19 | |
| F-PointNetModality=RGB+LiDAR2020.12 | 91.17 | 84.67 | 74.77 | |
| AVOD-FPNModality=RGB+LiDAR2020.12 | 90.99 | 84.82 | 79.62 | |
| Fast Point R-CNNModality=LiDAR only2020.12 | 90.87 | 87.84 | 80.52 | |
| PointPillarsModality=LiDAR only2020.12 | 90.07 | 86.56 | 82.81 | |
| PointPillarsMethod Architecture=One-stage, Latency (ms)=242026.02 | 90.07 | 86.56 | 82.81 | |
| SECONDMethod Architecture=One-stage, Latency (ms)=502026.02 | 89.39 | 83.77 | 78.59 | |
| VoxelNetModality=LiDAR only2020.12 | 89.35 | 79.26 | 77.39 | |
| SECONDModality=LiDAR only2020.12 | 88.07 | 79.37 | 77.95 | |
| MV3DModality=RGB+LiDAR2020.12 | 86.62 | 78.93 | 69.8 | |
| OursTime=0.05s2021.02 | 29.81 | 17.98 | 13.08 | |
| AM3DTime=0.4s2021.02 | 25.03 | 17.32 | 14.91 | |
| AM3DExtra=Depth, IoU threshold=0.72021.03 | 25.03 | 17.32 | 14.91 | |
| M3DSSDExtra=None, IoU threshold=0.72021.03 | 24.15 | 15.93 | 12.11 | |
| D4LCNTime=0.2s2021.02 | 22.51 | 16.02 | 12.55 | |
| D4LCNExtra=Depth, IoU threshold=0.72021.03 | 22.51 | 16.02 | 12.55 | |
| PLiDARTime=0.1s2021.02 | 21.27 | 13.92 | 11.25 | |
| M3D-RPNTime=0.16s2021.02 | 21.02 | 13.67 | 10.42 | |
| M3D-RPNExtra=None, IoU threshold=0.72021.03 | 21.02 | 13.67 | 10.23 | |
| SMOKEExtra=None, IoU threshold=0.72021.03 | 20.83 | 14.49 | 12.75 | |
| MonoPairExtra=None, IoU threshold=0.72021.03 | 19.28 | 14.83 | 12.89 | |
| RTM3DTime=0.05s2021.02 | 19.17 | 14.2 | 11.99 | |
| RTM3DExtra=None, IoU threshold=0.72021.03 | 19.17 | 14.2 | 11.99 | |
| MonoPSRTime=0.2s2021.02 | 18.33 | 12.58 | 9.91 | |
| MonoPSRExtra=None, IoU threshold=0.72021.03 | 18.33 | 12.58 | 9.91 | |
| MonoGRNetExtra=None, IoU threshold=0.72021.03 | 18.19 | 11.17 | 8.73 | |
| MonoDISTime=0.1s2021.02 | 17.23 | 13.19 | 11.12 | |
| MonoDISExtra=None, IoU threshold=0.72021.03 | 17.23 | 13.19 | 11.12 | |
| MonoFENetExtra=Depth, IoU threshold=0.72021.03 | 17.03 | 11.03 | 9.05 | |
| SS3DTime=0.05s2021.02 | 16.33 | 11.52 | 9.93 | |
| SS3DExtra=None, IoU threshold=0.72021.03 | 16.33 | 11.52 | 9.93 | |
| GS3DExtra=None, IoU threshold=0.72021.03 | 8.41 | 6.08 | 4.94 |