3D Object Detection on nuScenes v1.0 (test)
77.9mAPMV2DFusion-E
Evaluation results
| Method | Links | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MV2DFusion-EModality=C+L, Ensemble & TTA=true2026.02 | 77.9 | 78.8 | — | — | — | — | — | — | — | — | — | — | 0.237 | 0.226 | 0.247 | 0.192 | 0.119 | |
| BEVFusion4D-EModality=C+L, Ensemble & TTA=true2026.02 | 76.8 | 77.2 | — | — | — | — | — | — | — | — | — | — | 0.229 | 0.229 | 0.302 | 0.225 | 0.135 | |
| EA-LSS-EModality=C+L, Ensemble & TTA=true2026.02 | 76.6 | 77.6 | — | — | — | — | — | — | — | — | — | — | 0.234 | 0.228 | 0.278 | 0.204 | 0.124 | |
| CMT-EModality=C+L, Ensemble & TTA=true2026.02 | 75.3 | 77 | — | — | — | — | — | — | — | — | — | — | 0.233 | 0.22 | 0.271 | 0.212 | 0.127 | |
| BEVFusion-EModality=C+L, Ensemble & TTA=true2026.02 | 75 | 76.1 | — | — | — | — | — | — | — | — | — | — | 0.242 | 0.227 | 0.32 | 0.222 | 0.13 | |
| MambaFusionModality=C+L, Ensemble & TTA=false2026.02 | 74.7 | 77.2 | — | — | — | — | — | — | — | — | — | — | 0.248 | 0.232 | 0.291 | 0.221 | 0.118 | |
| MV2DFusionModality=C+L, Ensemble & TTA=false2026.02 | 74.5 | 76.7 | — | — | — | — | — | — | — | — | — | — | 0.245 | 0.229 | 0.269 | 0.199 | 0.115 | |
| BEVFusion4DModality=C+L, Ensemble & TTA=false2026.02 | 73.3 | 74.7 | — | — | — | — | — | — | — | — | — | — | 0.252 | 0.237 | 0.315 | 0.248 | 0.127 | |
| FocalFormer3D-FModality=LiDAR+Camera, flipping test-time augmentation=true2023.08 | 72.9 | 75 | 88.8 | 63.5 | 73.7 | 66.9 | 39 | 91.1 | 81 | 58.1 | 87.1 | 79.2 | — | — | — | — | — | |
| UniMTModality=L+C2025.04 | 72.7 | 75.3 | — | — | — | — | — | — | — | — | — | — | 23.9 | 23.4 | 27.4 | 23.3 | 12.7 | |
| FusionFormerModality=C+L, Ensemble & TTA=false2026.02 | 72.6 | 75.1 | — | — | — | — | — | — | — | — | — | — | 0.267 | 0.236 | 0.286 | 0.225 | 0.105 | |
| FusionFormerModality=L+C+T2025.04 | 72.6 | 75.1 | — | — | — | — | — | — | — | — | — | — | 26.7 | 23.6 | 28.6 | 22.5 | 10.5 | |
| FocalFormer3D-FModality=L+C2025.04 | 72.4 | 74.5 | — | — | — | — | — | — | — | — | — | — | 25.1 | 24.2 | 32.8 | 22.6 | 12.6 | |
| GaussianFusionModality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 72.4 | 74.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| EA-LSSModality=C+L, Ensemble & TTA=false2026.02 | 72.2 | 74.4 | — | — | — | — | — | — | — | — | — | — | 0.247 | 0.237 | 0.304 | 0.25 | 0.133 | |
| EA-LSSModality=L+C2025.04 | 72.2 | 74.4 | — | — | — | — | — | — | — | — | — | — | 24.7 | 23.7 | 30.4 | 25 | 13.3 | |
| EA-LSSModality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 72.2 | 74.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CMTModality=C+L, Ensemble & TTA=false2026.02 | 72 | 74.1 | — | — | — | — | — | — | — | — | — | — | 0.279 | 0.235 | 0.308 | 0.259 | 0.112 | |
| CMTModality=L+C2025.04 | 72 | 74.1 | — | — | — | — | — | — | — | — | — | — | 27.9 | 23.5 | 30.8 | 25.9 | 11.2 | |
| DALModality=L+C2025.04 | 72 | 74.8 | — | — | — | — | — | — | — | — | — | — | 25.3 | 23.8 | 33.4 | 17.4 | 12 | |
| FocalFormer3DModality=LiDAR+Camera2023.08 | 71.6 | 73.9 | 88.5 | 61.4 | 71.7 | 66.4 | 35.9 | 89.7 | 80.3 | 57.1 | 85.3 | 79.3 | — | — | — | — | — | |
| MSMDFusionModality=L+C2025.04 | 71.5 | 74 | — | — | — | — | — | — | — | — | — | — | 25.5 | 23.8 | 31 | 24.4 | 13.2 | |
| MSMDFusionModality=C+L, Resolution=800×448, Backbone Camera=ResNet-50, Backbone Lidar=VoxelNet2026.07 | 71.5 | 74 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BEVFusionModality=C+L, Ensemble & TTA=false, config_variant=higher_performance_row2026.02 | 71.3 | 73.3 | — | — | — | — | — | — | — | — | — | — | 0.25 | 0.24 | 0.359 | 0.254 | 0.132 | |
| LargeKernel3D-FModality=LiDAR+Camera, flipping and rotation test-time augmentation=true2023.08 | 71.1 | 74.2 | 88.1 | 60.3 | 69.1 | 66.5 | 34.3 | 89.6 | 82 | 60.3 | 85.7 | 75.5 | — | — | — | — | — | |
| UniTRModality=L+C2025.04 | 70.9 | 74.5 | — | — | — | — | — | — | — | — | — | — | 24.1 | 22.9 | 25.6 | 24 | 13.1 | |
| UniTRModality=C+L, Resolution=704×256, Backbone Camera=-, Backbone Lidar=-2026.07 | 70.9 | 74.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeepInteractionModality=LiDAR+Camera2023.08 | 70.8 | 73.4 | 87.9 | 60.2 | 70.8 | 63.8 | 37.5 | 91.7 | 75.4 | 54.5 | 87.2 | 80.4 | — | — | — | — | — | |
| DeepInteractionModality=C+L, Ensemble & TTA=false2026.02 | 70.8 | 73.4 | — | — | — | — | — | — | — | — | — | — | 0.257 | 0.24 | 0.325 | 0.245 | 0.128 | |
| FocalFormer3DModality=LiDAR, flipping test-time augmentation=true2023.08 | 70.5 | 73.9 | 87.8 | 59.4 | 73 | 65.7 | 37.8 | 90 | 77.4 | 52.4 | 83.4 | 77.8 | — | — | — | — | — | |
| BEVFusion-MITModality=LiDAR+Camera2023.08 | 70.2 | 72.9 | 88.6 | 60.1 | 69.8 | 63.8 | 39.3 | 89.2 | 74.1 | 51 | 86.5 | 80 | — | — | — | — | — | |
| BEVFusionModality=C+L, Ensemble & TTA=false2026.02 | 70.2 | 72.9 | — | — | — | — | — | — | — | — | — | — | 0.261 | 0.239 | 0.329 | 0.26 | 0.134 | |
| BEVFusionModality=L+C2025.04 | 70.2 | 72.9 | — | — | — | — | — | — | — | — | — | — | 26.1 | 23.9 | 32.9 | 26 | 13.4 | |
| BEVFusion(M)Modality=C+L, Resolution=704×256, Backbone Camera=Swin-T, Backbone Lidar=VoxelNet2026.07 | 70.2 | 72.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Focals Conv-FModality=LiDAR+Camera, flipping and rotation test-time augmentation=true2023.08 | 70.1 | 73.6 | 87.5 | 60 | 69.9 | 64 | 32.6 | 89 | 81.1 | 59.2 | 85.5 | 71.8 | — | — | — | — | — | |
| Fore-Mamba3DCBGS=false2026.02 | 70.1 | 74 | 87.1 | 60.6 | 70.5 | 63.9 | 34.3 | 90.2 | 73.7 | 53.1 | 88.6 | 78.6 | — | — | — | — | — | |
| LIONPresent at=NIPS24, CBGS=false2026.02 | 69.8 | 73.9 | 87.2 | 61.1 | 68.9 | 65 | 36.3 | 90 | 74 | 49.2 | 87.3 | 79.5 | — | — | — | — | — | |
| BEVFusionModality=LiDAR+Camera2023.08 | 69.2 | 71.8 | 88.1 | 60.9 | 69.3 | 62.1 | 34.4 | 89.2 | 72.2 | 52.2 | 85.2 | 78.2 | — | — | — | — | — | |
| Voxel-MambaPresent at=NIPS24, CBGS=false2026.02 | 69 | 73 | 86.8 | 57.1 | 68 | 63.2 | 35.4 | 89.5 | 74.7 | 50.8 | 86.9 | 77.3 | — | — | — | — | — | |
| TransFusionModality=LiDAR+Camera2023.08 | 68.9 | 71.7 | 87.1 | 60 | 68.3 | 60.8 | 33.1 | 88.4 | 73.6 | 52.9 | 86.7 | 78.1 | — | — | — | — | — | |
| TransFusionModality=C+L, Ensemble & TTA=false2026.02 | 68.9 | 71.7 | — | — | — | — | — | — | — | — | — | — | 0.259 | 0.243 | 0.359 | 0.288 | 0.127 | |
| TransFusionModality=L+C2025.04 | 68.9 | 71.7 | — | — | — | — | — | — | — | — | — | — | 25.9 | 24.3 | 35.9 | 28.8 | 12.7 | |
| LargeKernel3DModality=LiDAR, flipping test-time augmentation=true2023.08 | 68.8 | 72.8 | 87.3 | 59.1 | 68.5 | 65.6 | 30.2 | 88.3 | 77.8 | 53.5 | 82.4 | 75 | — | — | — | — | — | |
| FocalFormer3DModality=LiDAR2023.08 | 68.7 | 72.6 | 87.2 | 57.1 | 69.6 | 64.9 | 34.4 | 88.2 | 76.2 | 49.6 | 82.3 | 77.8 | — | — | — | — | — | |
| AutoAlignV2Modality=L+C2025.04 | 68.4 | 72.4 | — | — | — | — | — | — | — | — | — | — | 24.5 | 23.3 | 31.1 | 25.8 | 13.3 | |
| AutoAlignV2Modality=C+L, Resolution=1600×640, Backbone Camera=CSPNet, Backbone Lidar=VoxelNet2026.07 | 68.4 | 72.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAFDNetPresent at=CVPR24, CBGS=false2026.02 | 68.3 | 72.3 | 87.3 | 57.3 | 68 | 63.7 | 37.3 | 89 | 71.1 | 44.8 | 84.9 | 79.5 | — | — | — | — | — | |
| FusionPaintingModality=LiDAR+Camera2023.08 | 68.1 | 71.6 | 87.1 | 60.8 | 68.5 | 61.7 | 30 | 88.3 | 74.7 | 53.5 | 85 | 71.8 | — | — | — | — | — | |
| HEDNet2023.10 | 67.7 | 72 | 87.1 | 56.5 | 70.4 | 63.5 | 33.6 | 87.9 | 70.4 | 44.8 | 85.1 | 78.1 | — | — | — | — | — | |
| HEDNetPresent at=NIPS24, CBGS=false2026.02 | 67.7 | 72 | 87.1 | 56.5 | 70.4 | 63.5 | 33.6 | 87.9 | 70.4 | 44.8 | 85.1 | 78.1 | — | — | — | — | — | |
| MGTANetModality=LiDAR, flipping test-time augmentation=true2023.08 | 67.5 | 72.7 | 88.5 | 59.8 | 67.2 | 61.5 | 30.6 | 87.3 | 75.8 | 52.5 | 85.5 | 66.3 | — | — | — | — | — | |
| UVTRModality=C+L, Ensemble & TTA=false2026.02 | 67.1 | 71.1 | — | — | — | — | — | — | — | — | — | — | 0.306 | 0.245 | 0.351 | 0.225 | 0.124 | |
| LiDARMultiNetModality=LiDAR2023.08 | 67 | 71.6 | 86.9 | 57.4 | 64.7 | 61 | 31.5 | 87.2 | 75.3 | 47.6 | 85.1 | 73.5 | — | — | — | — | — | |
| PointAugmentingModality=LiDAR+Camera, flipping test-time augmentation=true2023.08 | 66.8 | 71 | 87.5 | 57.3 | 65.2 | 60.7 | 28 | 87.9 | 74.3 | 50.9 | 83.6 | 72.6 | — | — | — | — | — | |
| MVPModality=LiDAR+Camera2023.08 | 66.4 | 70.5 | 86.8 | 58.5 | 67.4 | 57.3 | 26.1 | 89.1 | 70 | 49.3 | 85 | 74.8 | — | — | — | — | — | |
| MVPModality=C+L, Ensemble & TTA=false2026.02 | 66.4 | 70.5 | — | — | — | — | — | — | — | — | — | — | 0.263 | 0.238 | 0.321 | 0.313 | 0.134 | |
| LinK2023.10 | 66.3 | 71 | 86.1 | 55.7 | 65.7 | 62.1 | 30.9 | 85.8 | 73.5 | 47.5 | 80.4 | 75.5 | — | — | — | — | — | |
| PillarNet-34Modality=LiDAR2023.08 | 66 | 71.4 | 87.6 | 57.5 | 63.6 | 63.1 | 27.9 | 87.3 | 70.1 | 42.3 | 83.3 | 77.2 | — | — | — | — | — | |
| TransFusion-LModality=LiDAR2023.08 | 65.5 | 70.2 | 86.2 | 56.7 | 66.3 | 58.8 | 28.2 | 86.1 | 68.3 | 44.2 | 82 | 78.2 | — | — | — | — | — | |
| TransFusion-L2023.10 | 65.5 | 70.2 | 86.2 | 56.7 | 66.3 | 58.8 | 28.2 | 86.1 | 68.3 | 44.2 | 82 | 78.2 | — | — | — | — | — | |
| LargeKernel3D2023.10 | 65.4 | 70.6 | 85.5 | 53.8 | 64.4 | 59.5 | 29.7 | 85.9 | 72.7 | 46.8 | 79.9 | 75.5 | — | — | — | — | — | |
| LargeKernel3DModality=LiDAR2023.08 | 65.3 | 70.5 | 85.9 | 55.3 | 66.2 | 60.2 | 26.8 | 85.6 | 72.5 | 46.6 | 80 | 74.3 | — | — | — | — | — | |
| RaCFormerInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=24, Temporal frames (+t)=62024.12 | 64.9 | 70.2 | — | — | — | — | — | — | — | — | — | — | 0.358 | 0.24 | 0.329 | 0.179 | 0.119 | |
| VoxelNeXt2023.10 | 64.5 | 70 | 84.6 | 53 | 64.7 | 55.8 | 28.7 | 85.8 | 73.2 | 45.7 | 79 | 74.6 | — | — | — | — | — | |
| VoxelNeXtInput=L, Backbone=Sparse CNNs, Epochs=202024.12 | 64.5 | 70 | — | — | — | — | — | — | — | — | — | — | 0.268 | 0.238 | 0.377 | 0.219 | 0.127 | |
| UVTR-L2023.10 | 63.9 | 69.7 | 86.3 | 52.2 | 62.8 | 59.7 | 33.7 | 84.5 | 68.8 | 41.1 | 74.7 | 74.9 | — | — | — | — | — | |
| Focals Conv2023.10 | 63.8 | 70 | 86.7 | 56.3 | 67.7 | 59.5 | 23.8 | 87.5 | 64.5 | 36.3 | 81.4 | 74.1 | — | — | — | — | — | |
| Far3DBackbone=ViT-L, Resolution=1536 x 15362023.08 | 63.5 | 68.7 | — | — | — | — | — | — | — | — | — | — | 0.432 | 0.237 | 0.278 | 0.227 | 0.13 | |
| VISTA-OHSruntime=69ms2022.03 | 63 | 69.8 | 84.4 | 55.1 | 63.7 | 54.2 | 25.1 | 82.8 | 70 | 45.4 | 78.5 | 71.4 | — | — | — | — | — | |
| VISTA2023.10 | 63 | 69.8 | 84.4 | 55.1 | 63.7 | 54.2 | 25.1 | 82.8 | 70 | 45.4 | 78.5 | 71.4 | — | — | — | — | — | |
| AFDetV22023.10 | 62.4 | 68.5 | 86.3 | 54.2 | 62.5 | 58.9 | 26.7 | 85.8 | 63.8 | 34.3 | 80.1 | 71 | — | — | — | — | — | |
| StreamPETRBackbone=ViT-L, Resolution=1536 x 15362023.08 | 62 | 67.6 | — | — | — | — | — | — | — | — | — | — | 0.47 | 0.241 | 0.258 | 0.236 | 0.134 | |
| ConFusionInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=242026.04 | 61.6 | 67.9 | — | — | — | — | — | — | — | — | — | — | 0.382 | 0.236 | 0.348 | 0.201 | 0.123 | |
| Multi-View Framework with Panoptic Guidance2022.03 | 60.9 | 67.3 | 84.6 | 50 | 63.2 | 55.3 | 23.4 | 83.7 | 65.1 | 38.9 | 76.8 | 68.2 | — | — | — | — | — | |
| HVDetFusionInput=C+R, Image Size=640×1600, Backbone=InternImage-B, Epochs=20, Temporal frames (+t)=82024.12 | 60.9 | 67.4 | — | — | — | — | — | — | — | — | — | — | 0.379 | 0.243 | 0.382 | 0.172 | 0.132 | |
| CenterPointruntime=70ms2022.03 | 60.3 | 67.3 | 85.2 | 53.5 | 63.6 | 56 | 20 | 84.6 | 59.5 | 30.7 | 78.4 | 71.1 | — | — | — | — | — | |
| CenterPointModality=LiDAR, flipping test-time augmentation=true2023.08 | 60.3 | 67.3 | 85.2 | 53.5 | 63.6 | 56 | 20 | 84.6 | 59.5 | 30.7 | 78.4 | 71.1 | — | — | — | — | — | |
| CenterPointInput=L, Backbone=VoxelNet, Epochs=202024.12 | 60.3 | 67.3 | — | — | — | — | — | — | — | — | — | — | 0.262 | 0.239 | 0.361 | 0.288 | 0.136 | |
| FCOS-LiDAR2023.10 | 60.2 | 65.7 | 82.2 | 47.7 | 63.9 | 48.8 | 28.8 | 84.5 | 68 | 39 | 79.2 | 70.7 | — | — | — | — | — | |
| SpaRCInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=602026.04 | 60 | 67.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HotSpotNet2022.03 | 59.3 | 66 | 83.1 | 50.9 | 56.4 | 53.3 | 23 | 81.3 | 63.5 | 36.6 | 73 | 71.6 | — | — | — | — | — | |
| OHSruntime=60ms2022.03 | 59.3 | 66 | 83.1 | 50.9 | 56.4 | 53.3 | 23 | 81.3 | 63.5 | 36.6 | 73 | 71.6 | — | — | — | — | — | |
| HotSpotNet2023.10 | 59.3 | 66 | 83.1 | 50.9 | 56.4 | 53.3 | 23 | 81.3 | 63.5 | 36.6 | 73 | 71.6 | — | — | — | — | — | |
| RaCFormerInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=242024.12 | 59.2 | 65.9 | — | — | — | — | — | — | — | — | — | — | 0.407 | 0.244 | 0.345 | 0.238 | 0.132 | |
| RaCFormerInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=242026.04 | 59.2 | 65.9 | — | — | — | — | — | — | — | — | — | — | 0.407 | 0.244 | 0.345 | 0.238 | 0.132 | |
| CRT-FusionInput=C+R, Image Size=640×1600, Backbone=ConvNeXt-B, Epochs=242026.04 | 58.3 | 64.9 | — | — | — | — | — | — | — | — | — | — | 0.365 | 0.261 | 0.405 | 0.262 | 0.132 | |
| CVCNET2023.10 | 58.2 | 66.6 | 82.6 | 49.5 | 59.4 | 51.1 | 16.2 | 83 | 61.8 | 38.8 | 69.7 | 69.7 | — | — | — | — | — | |
| CenterPoint2022.03 | 58 | 65.5 | 84.6 | 51 | 60.2 | 53.2 | 17.5 | 83.4 | 53.7 | 28.7 | 76.7 | 70.9 | — | — | — | — | — | |
| CenterPoint-VoxelModality=LiDAR2022.03 | 58 | 65.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CenterPoint2023.10 | 58 | 65.5 | 84.6 | 51 | 60.2 | 53.2 | 17.5 | 83.4 | 53.7 | 28.7 | 76.7 | 70.9 | — | — | — | — | — | |
| CRNInput=C+R, Backbone=ConvNeXt-B2024.03 | 57.5 | 62.4 | — | — | — | — | — | — | — | — | — | — | 0.416 | 0.264 | 0.456 | 0.365 | 0.13 | |
| CRNInput=C+R, Image Size=640×1600, Backbone=ConvNeXt-B, Epochs=242024.12 | 57.5 | 62.4 | — | — | — | — | — | — | — | — | — | — | 0.416 | 0.264 | 0.456 | 0.365 | 0.13 | |
| CRNInput=C+R, Backbone=ConvNeXt-B2024.03 | 57.5 | 62.4 | — | — | — | — | — | — | — | — | — | — | 0.416 | — | — | — | — | |
| CRNBackbone=V2-99, Image Size=640 × 1600, Input=Camera + Radar2026.04 | 57.5 | 62.4 | — | — | — | — | — | — | — | — | — | — | 0.416 | 0.264 | 0.456 | 0.365 | 0.13 | |
| CRNInput=C+R, Image Size=640×1600, Backbone=ConvNeXt-B, Epochs=242026.04 | 57.5 | 62.4 | — | — | — | — | — | — | — | — | — | — | 0.416 | 0.264 | 0.456 | 0.365 | 0.13 | |
| HyDRaInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=202024.12 | 57.4 | 64.2 | — | — | — | — | — | — | — | — | — | — | 0.398 | 0.251 | 0.423 | 0.249 | 0.122 | |
| HyDRaInput=C+R, Backbone=V2-992024.03 | 57.4 | 64.2 | — | — | — | — | — | — | — | — | — | — | 0.398 | — | — | — | — | |
| HyDRaInput=C+R, Image Size=640×1600, Backbone=V2-99, Epochs=202026.04 | 57.4 | 64.2 | — | — | — | — | — | — | — | — | — | — | 0.398 | 0.251 | 0.423 | 0.249 | 0.122 | |
| CAM3DNetBackbone=V2-99, Image Size=640 × 1600, Input=Camera2026.04 | 57.09 | 64.13 | — | — | — | — | — | — | — | — | — | — | 0.4575 | 0.2419 | 0.3462 | 0.2764 | 0.1201 | |
| CVCNetruntime=91ms2022.03 | 55.8 | 64.2 | 82.7 | 46.1 | 45.8 | 46.7 | 20.7 | 81 | 61.3 | 34.3 | 69.7 | 69.9 | — | — | — | — | — |