3D Object Detection (Cyclist) on Waymo Open Dataset (WOD) v1.2 (val)
72.58mAP L1OcTr
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| OcTrTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 72.58 | 71.5 | 69.93 | 68.9 | |
| PV-RCNN++Training samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 71.8 | 70.71 | 69.31 | 68.26 | |
| PDVTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 68.71 | 67.55 | 66.49 | 65.36 | |
| CenterPointTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 68.68 | 67.39 | 66.11 | 64.87 | |
| LiDAR-RCNNTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 68.6 | 66.9 | 66.1 | 64.4 | |
| PartA2NetTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 67.35 | 66.15 | 65.05 | 63.89 | |
| PV-RCNNTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 65.88 | 64.25 | 63.39 | 61.82 | |
| SECONDTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 57.13 | 55.62 | 54.97 | 53.53 | |
| PointPillarTraining samples ratio=20%, IoU threshold=0.5, Training mode=Multi-class2023.03 | 55.26 | 51.75 | 53.18 | 49.8 |