3D Object Detection on nuScenes v1.0-trainval (val)
74NDSGaussianFusion
Evaluation Results
| Method | Links | ||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GaussianFusionModality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 74 | — | — | 71.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MEFormerDrop rate=0%2026.03 | 73.86 | — | — | 71.49 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniTRModality=C+L, Resolution=704×256, Backbone Camera=-, Backbone Lidar=-2026.07 | 73.3 | — | — | 70.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FusionFormer-SModality=C+L, Resolution=1600×640, Backbone Camera=VoV-99, Backbone Lidar=VoxelNet2026.07 | 73.2 | — | — | 70 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| EA-LSSModality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 73.1 | — | — | 71.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMTTime offset=0s2026.01 | 72.9 | — | — | 70.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69 | 61 | 47.5 | 36.8 | — | — | — | |
| CMT + EMCTime offset=0s2026.01 | 72.9 | — | — | 70.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 69 | 61 | 47.5 | 36.8 | — | — | — | |
| LiteFusion-L3D Backbone=No, 2D Backbone=ConvNext-B2025.12 | 72.9 | — | — | 67.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMT + AsyncBEVTime offset=0s2026.01 | 72.7 | — | — | 69.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 68.8 | 60.8 | 48.1 | 37 | — | — | — | |
| CMT + AsyncBEVTime offset=0.25s2026.01 | 72.7 | — | — | 69.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 68.8 | 60.6 | 48.1 | 36.9 | — | — | — | |
| CMTDrop rate=0%2026.03 | 72.68 | — | — | 70.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMT + AsyncBEVTime offset=0.5s2026.01 | 72.3 | — | — | 69.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 68.4 | 60.2 | 48 | 36.7 | — | — | — | |
| Fore-Mamba3DCBGS=false2026.02 | 72.3 | — | — | 68.4 | — | — | — | — | — | 88.4 | 65.2 | 80.3 | 48 | 28.2 | 89.3 | 75.7 | 57.7 | 80 | 71.2 | — | — | — | — | — | — | — | |
| MEFormer + ModalPatchDrop rate=10%2026.03 | 72.11 | — | — | 68.65 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LIONPresent at=NIPS24, CBGS=false2026.02 | 72.1 | — | — | 68 | — | — | — | — | — | 87.9 | 64.9 | 77.6 | 44.4 | 28.5 | 89.6 | 75.6 | 59.4 | 80.8 | 71.6 | — | — | — | — | — | — | — | |
| MSMDFusionModality=C+L, Resolution=800×448, Backbone Camera=ResNet-50, Backbone Lidar=VoxelNet2026.07 | 72.1 | — | — | 69.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FASD(Student)Param=6.13M, Flops=54.7G, Memory=546 MiB, FPS=19.462024.09 | 71.98 | — | — | 66.98 | 0.2732 | 0.2512 | 0.2657 | 0.262 | 0.183 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Voxel-MambaPresent at=NIPS24, CBGS=false2026.02 | 71.9 | — | — | 67.5 | — | — | — | — | — | 87.9 | 62.8 | 76.8 | 45.9 | 24.9 | 89.3 | 77.1 | 58.6 | 80.1 | 71.5 | — | — | — | — | — | — | — | |
| FSHNetbasetraining schedule=36 epochs without CBGS, re-implemented=true2025.06 | 71.7 | — | — | 68.1 | — | — | — | — | — | 88.7 | 61.4 | 79.3 | 47.8 | 26.3 | 89.3 | 76.7 | 60.5 | 78.6 | 72.3 | — | — | — | — | — | — | — | |
| FSHNetPresent at=CVPR25, CBGS=false2026.02 | 71.7 | — | — | 68.1 | — | — | — | — | — | 88.7 | 61.4 | 79.3 | 47.8 | 26.3 | 89.3 | 76.7 | 60.5 | 78.6 | 72.3 | — | — | — | — | — | — | — | |
| MetaBEV3D Backbone=Yes, 2D Backbone=SwinT2025.12 | 71.5 | — | — | 68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MetaBEVModality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 71.5 | — | — | 68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VoxelMambaParam=6.73M, Flops=58.5G, Memory=653 MiB, FPS=11.272024.09 | 71.48 | — | — | 67.54 | 0.2785 | 0.2583 | 0.2584 | 0.2687 | 0.1899 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BevFusionDrop rate=0%2026.03 | 71.47 | — | — | 68.73 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMT + ModalPatchDrop rate=10%2026.03 | 71.42 | — | — | 67.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMT + EMCTime offset=0.25s2026.01 | 71.4 | — | — | 67.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 67.6 | 58.9 | 46.9 | 36 | — | — | — | |
| BEVFusion (MIT)3D Backbone=Yes, 2D Backbone=SwinT2025.12 | 71.4 | — | — | 68.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HEDNetPresent at=NIPS23, CBGS=false2026.02 | 71.4 | — | — | 66.7 | — | — | — | — | — | 87.7 | 60.6 | 77.8 | 50.7 | 28.9 | 87.1 | 74.3 | 56.8 | 76.3 | 66.9 | — | — | — | — | — | — | — | |
| BEVFusion(M)Modality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 71.4 | — | — | 68.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AutoAlignV23D Backbone=Yes, 2D Backbone=CSPNet2025.12 | 71.2 | — | — | 67.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AutoAlignV2Modality=C+L, Resolution=1600×640, Backbone Camera=CSPNet, Backbone Lidar=VoxelNet2026.07 | 71.2 | — | — | 67.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DSVT2025.06 | 71.1 | — | — | 66.4 | — | — | — | — | — | 87.4 | 62.6 | 75.9 | 42.1 | 25.3 | 88.2 | 74.8 | 58.7 | 77.8 | 70.9 | — | — | — | — | — | — | — | |
| VCD-E3D Backbone=No, 2D Backbone=ConvNext-B2025.12 | 71.1 | — | — | 67.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DSVTPresent at=CVPR23, CBGS=false2026.02 | 71.1 | — | — | 66.4 | — | — | — | — | — | 87.4 | 62.6 | 75.9 | 42.1 | 25.3 | 88.2 | 74.8 | 58.7 | 77.9 | 71 | — | — | — | — | — | — | — | |
| SAFDNetre-implemented=true2025.06 | 71 | — | — | 66.3 | — | — | — | — | — | 87.6 | 60.8 | 78 | 43.5 | 26.6 | 87.8 | 75.5 | 58 | 75 | 69.7 | — | — | — | — | — | — | — | |
| SAFDNetPresent at=CVPR24, CBGS=false2026.02 | 71 | — | — | 66.3 | — | — | — | — | — | 87.6 | 60.8 | 78 | 43.5 | 26.6 | 87.8 | 75.5 | 58 | 75 | 69.7 | — | — | — | — | — | — | — | |
| CMT3D Backbone=Yes, 2D Backbone=ResNet-502025.12 | 70.8 | — | — | 67.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMTTime offset=0.25s2026.01 | 70.7 | — | — | 66.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 66.9 | 57.9 | 46.9 | 36 | — | — | — | |
| GraphAlign3D Backbone=Yes, 2D Backbone=ResNet-1012025.12 | 70.6 | — | — | 66.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FASD(Teacher)Param=9.12M, Flops=121.7G, Memory=4085 MiB, FPS=9.152024.09 | 70.53 | — | — | 66.58 | 0.2732 | 0.254 | 0.3024 | 0.2777 | 0.1821 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FSDv2re-implemented=true2025.06 | 70.4 | — | — | 64.7 | — | — | — | — | — | 84.4 | 57.3 | 75.9 | 44.1 | 28.5 | 86.9 | 69.5 | 57.4 | 72.9 | 73.6 | — | — | — | — | — | — | — | |
| BEVFusion† (PKU)3D Backbone=Yes, 2D Backbone=Dual-Swin-T2025.12 | 70.4 | — | — | 66.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UVTR3D Backbone=Yes, 2D Backbone=ResNet-1012025.12 | 70.2 | — | — | 65.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TransFusion3D Backbone=Yes, 2D Backbone=DLA342025.12 | 70.2 | — | — | 66.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TransFusion-L2025.06 | 70.1 | — | — | 65.5 | — | — | — | — | — | 86.9 | 60.8 | 73.1 | 43.4 | 25.2 | 87.5 | 72.9 | 57.3 | 77.2 | 70.3 | — | — | — | — | — | — | — | |
| TransFusion-LPresent at=CVPR22, CBGS=false2026.02 | 70.1 | — | — | 65.5 | — | — | — | — | — | 86.9 | 60.8 | 73.1 | 43.4 | 25.2 | 87.5 | 72.9 | 57.3 | 77.2 | 70.3 | — | — | — | — | — | — | — | |
| VoxelNeXtPresent at=CVPR23, CBGS=false2026.02 | 70 | — | — | 64.5 | — | — | — | — | — | 84.6 | 53 | 64.7 | 55.8 | 28.7 | 85.8 | 73.2 | 45.7 | 79 | 74.6 | — | — | — | — | — | — | — | |
| PointPainting3D Backbone=Yes, 2D Backbone=ResNet-1012025.12 | 69.9 | — | — | 65.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MEFormerDrop rate=10%2026.03 | 69.4 | — | — | 62.93 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DSVTParam=8.65M, Flops=110.2G, Memory=3966 MiB, FPS=9.582024.09 | 68.94 | — | — | 64.22 | 0.2877 | 0.2611 | 0.3701 | 0.2087 | 0.1851 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LiteFusion-B3D Backbone=No, 2D Backbone=ResNet-1012025.12 | 68.9 | — | — | 62.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMT + EMCTime offset=0.5s2026.01 | 68.8 | — | — | 63.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 65.1 | 54.9 | 45.9 | 34.1 | — | — | — | |
| VoxelNeXtre-implemented=true2025.06 | 68.7 | — | — | 63.5 | — | — | — | — | — | 83.9 | 55.5 | 70.5 | 38.1 | 21.1 | 84.6 | 62.8 | 50 | 69.4 | 69.4 | — | — | — | — | — | — | — | |
| UniBEVDrop rate=0%2026.03 | 68.53 | — | — | 64.24 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEV3D Backbone=Yes, 2D Backbone=ResNet-1012025.12 | 68.5 | — | — | 64.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BevFusion + ModalPatchDrop rate=10%2026.03 | 68.43 | — | — | 63.37 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PillarNeXt2025.06 | 68.4 | — | — | 62.2 | — | — | — | — | — | 85 | 57.4 | 67.6 | 35.6 | 20.6 | 86.8 | 68.6 | 53.1 | 77.3 | 69.7 | — | — | — | — | — | — | — | |
| FUTR3D3D Backbone=Yes, 2D Backbone=ResNet-1012025.12 | 68.3 | — | — | 64.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMTDrop rate=10%2026.03 | 67.81 | — | — | 60.51 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMTTime offset=0.5s2026.01 | 67.5 | — | — | 61.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.8 | 52.4 | 45.8 | 34.2 | — | — | — | |
| BevFusionDrop rate=10%2026.03 | 67.46 | — | — | 60.74 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEV + ModalPatchDrop rate=10%2026.03 | 67.35 | — | — | 62.65 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VoxelNextParam=19.05M, Flops=57.1G, Memory=1970 MiB, FPS=18.122024.09 | 67.09 | — | — | 60.55 | 0.3023 | 0.2526 | 0.3701 | 0.2087 | 0.1851 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEV + AsyncBEVTime offset=0s2026.01 | 66.9 | — | — | 62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.4 | 51.7 | 42 | 32.3 | — | — | — | |
| UniBEV + AsyncBEVTime offset=0.25s2026.01 | 66.8 | — | — | 61.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.3 | 51.6 | 42.1 | 32.3 | — | — | — | |
| UniBEVTime offset=0s2026.01 | 66.7 | — | — | 62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.5 | 52.2 | 42.4 | 32.9 | — | — | — | |
| UniBEV + EMCTime offset=0s2026.01 | 66.7 | — | — | 62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.5 | 52.2 | 42.4 | 32.9 | — | — | — | |
| UniBEV + AsyncBEVTime offset=0.5s2026.01 | 66.7 | — | — | 61.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.2 | 51.5 | 42.2 | 32.3 | — | — | — | |
| CenterPoint2025.06 | 66.5 | — | — | 59.2 | — | — | — | — | — | 84.9 | 57.4 | 70.7 | 38.1 | 16.9 | 85.1 | 59 | 42 | 69.8 | 68.3 | — | — | — | — | — | — | — | |
| UniBEV + EMCTime offset=0.25s2026.01 | 66.5 | — | — | 61.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.5 | 52 | 41.9 | 32.2 | — | — | — | |
| CenterPointPresent at=CVPR21, CBGS=false2026.02 | 66.5 | — | — | 59.2 | — | — | — | — | — | 84.9 | 57.4 | 70.7 | 38.1 | 16.9 | 85.1 | 59 | 42 | 69.8 | 68.3 | — | — | — | — | — | — | — | |
| MEFormer + ModalPatchDrop rate=30%2026.03 | 66.3 | — | — | 59.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CenterPointParam=7.75M, Flops=48.5G, Memory=2360 MiB, FPS=21.582024.09 | 66.29 | — | — | 58.77 | 0.2919 | 0.2566 | 0.3692 | 0.2081 | 0.1837 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEV + EMCTime offset=0.5s2026.01 | 66.2 | — | — | 60.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 63.2 | 51.6 | 41 | 30.9 | — | — | — | |
| CMT + ModalPatchDrop rate=30%2026.03 | 65.81 | — | — | 58.77 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEVDrop rate=10%2026.03 | 65.47 | — | — | 59.66 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CenterPoint-VInput=L, Backbone=Voxel2024.03 | 65.3 | — | — | 56.9 | 0.285 | 0.253 | 0.323 | 0.272 | 0.186 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniBEVTime offset=0.25s2026.01 | 64.8 | — | — | 59.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 61.9 | 50.1 | 41.8 | 32.3 | — | — | — | |
| SparseBEV (Teacher)Backbone=ResNet-101, Frames=15, Input size=512 x 14082025.12 | 63.8 | — | — | 55.1 | 0.493 | 0.265 | 0.275 | 0.163 | 0.181 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Density-aware Temperature ScalingBackbone=SECOND, Calibration Category=Density-aware, Calibration Level=Global2026.05 | 63.65 | — | — | 54.36 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 2.875 | 10.753 | |
| Density-aware Temperature ScalingBackbone=SECOND, Calibration Category=Density-aware, Calibration Level=Class-wise2026.05 | 63.65 | — | — | 54.36 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.518 | 7.053 | |
| DepthBackbone=SECOND, Calibration Category=Post-hoc, Calibration Level=Global2026.05 | 63.58 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 12.494 | 36.744 | |
| DepthBackbone=SECOND, Calibration Category=Post-hoc, Calibration Level=Class-wise2026.05 | 63.58 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 4.749 | 19.102 | |
| Temperature ScalingBackbone=SECOND, Calibration Category=Post-hoc, Calibration Level=Global2026.05 | 63.58 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 3.443 | 12.711 | |
| Temperature ScalingBackbone=SECOND, Calibration Category=Post-hoc, Calibration Level=Class-wise2026.05 | 63.58 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 1.533 | 7.281 | |
| Ensemble ScalingBackbone=SECOND, Calibration Category=Train-time2026.05 | 63.57 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 3.4 | 11.461 | |
| KL DivergenceBackbone=SECOND, Calibration Category=Train-time2026.05 | 63.57 | — | — | 54.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 4.692 | 13.205 | |
| UniBEV + ModalPatchDrop rate=30%2026.03 | 63.4 | — | — | 56.56 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Cal-DETRBackbone=SECOND, Calibration Category=Sample-based2026.05 | 63.17 | — | — | 53.57 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 34.234 | 95.525 | |
| UniBEVTime offset=0.5s2026.01 | 62.7 | — | — | 56.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 60 | 47.1 | 40.9 | 30.6 | — | — | — | |
| Deep EnsemblesBackbone=SECOND, Calibration Category=Sample-based2026.05 | 62.36 | — | — | 50.83 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 23.089 | 76.02 | |
| Monte-Carlo DropoutBackbone=SECOND, Calibration Category=Sample-based2026.05 | 62.3 | — | — | 51.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 31.62 | 87.506 | |
| HyDRaInput=C+R, Backbone=R101, Image Size=512 x 14082024.03 | 61.7 | — | — | 53.6 | 0.416 | 0.264 | 0.407 | 0.231 | 0.186 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BevFusion + ModalPatchDrop rate=30%2026.03 | 61.3 | — | — | 50.59 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MEFormerDrop rate=30%2026.03 | 60.29 | — | — | 45.37 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CAM3DNetBackbone=V2-99, Image Size=320 × 800, FLOPs(G)=824.88226, Params(M)=89.0662026.04 | 60.21 | — | — | 51.68 | 0.5601 | 0.2568 | 0.2938 | 0.2594 | 0.1932 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CAM3DNetBackbone=ResNet101, Image Size=512 × 1408, perspective-view pre-training=true, FLOPs(G)=781.905, Params(M)=61.9382026.04 | 60 | — | — | 51.24 | 0.5482 | 0.2631 | 0.326 | 0.2411 | 0.1826 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CenterPoint-PInput=L, Backbone=Pillars2024.03 | 59.8 | — | — | 49.4 | 0.32 | 0.262 | 0.377 | 0.334 | 0.198 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Sparse4D v2Backbone=ResNet101, Image Size=512 × 1408, k-means anchor initialization=true2026.04 | 59.4 | — | — | 50.5 | 0.548 | 0.268 | 0.348 | 0.239 | 0.184 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Far3DBackbone=ResNet101, Image Size=512 × 1408, perspective-view pre-training=true, FLOPs(G)=791.512, Params(M)=61.9452026.04 | 59.4 | — | — | 51 | 0.551 | 0.258 | 0.372 | 0.238 | 0.195 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |