3D Object Detection On Nuscenes Lidar Only
Métriques
NDS
NDS (val)
mAP
mAP (val)
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | NDS | NDS (val) | mAP | mAP (val) |
|---|---|---|---|---|
| lion-linear-group-rnn-for-3d-object-detection | 73.9 | 72.1 | 69.8 | 68.0 |
| transfusion-robust-lidar-camera-fusion-for-3d | 70.2 | 70.1 | 65.5 | 65.1 |
| pointpillars-fast-encoders-for-object | 45.3 | - | 30.5 | - |
| dsvt-dynamic-sparse-voxel-transformer-with | 72.7 | 71.1 | 68.4 | 66.4 |
| class-balanced-grouping-and-sampling-for | 63.3 | 62.3 | 52.8 | 50.6 |
| pillarnet-high-performance-pillar-based-3d | 71.4 | - | 66.0 | - |
| center-based-3d-object-detection-and-tracking | 67.3 | 66.8 | 60.3 | 59.6 |