Table 4 Performance comparison results of CenterTransFuser model on nuScenes test set
From: CenterTransFuser: radar point cloud and visual information fusion for 3D object detection
Baseline | Cro-trans | DT | Filter | DP | NDS | mAP | mATE | mASE | mAOE | mAVE | mAAE |
|---|---|---|---|---|---|---|---|---|---|---|---|
(%) | (%) | (%) | (%) | (%) | (%) | (%) | |||||
✓ | 0.438 | 0.319 | 0.654 | 0.289 | 0.566 | 0.573 | 0.120 | ||||
✓ | ✓ | 0.463 | 0.340 | 0.639 | 0.265 | 0.542 | 0.547 | 0.145 | |||
✓ | ✓ | 0.442 | 0.321 | 0.642 | 0.260 | 0.550 | 0.563 | 0.140 | |||
✓ | ✓ | ✓ | 0.459 | 0.336 | 0.637 | 0.258 | 0.542 | 0.559 | 0.139 | ||
✓ | ✓ | ✓ | ✓ | ✓ | 0.471 | 0.347 | 0.628 | 0.252 | 0.523 | 0.527 | 0.135 |