3D Object Detection On Dair V2X I
Metrics
AP|R40(easy)
AP|R40(hard)
AP|R40(moderate)
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | AP|R40(easy) | AP|R40(hard) | AP|R40(moderate) |
|---|---|---|---|
| monouni-a-unified-vehicle-and-infrastructure | 90.92 | 87.2 | 87.2 |
| pointpillars-fast-encoders-for-object | 63.1 | 54.0 | 54.0 |
| bevheight-a-robust-framework-for-vision-based | 77.8 | 65.9 | 65.8 |
| mvx-net-multimodal-voxelnet-for-3d-object | 71.0 | 53.8 | 53.7 |
| calibration-free-bev-representation-for | 72.0 | 60.1 | 60.1 |
| cobev-elevating-roadside-3d-object-detection | 82.0 | 69.7 | 69.6 |
| bevformer-learning-bird-s-eye-view | 61.4 | 50.7 | 50.7 |
| bevdepth-acquisition-of-reliable-depth-for | 75.7 | 63.7 | 63.6 |
| imvoxelnet-image-to-voxels-projection-for | 44.8 | 37.6 | 37.6 |