3D Object Detection on Waymo to KITTI Adaptation (W. -> K.)
91.19AP@0.7Target domain
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Target domainadaptation_type=Oracle/Upper Bound2025.12 | 91.19 | — | |
| 10 f. + L2-SPTarget Frames=10, Regularization=L2-SP2025.12 | 88.91 | 55.2 | |
| 10 f. + const LRTarget Frames=10, Learning Rate=constant2025.12 | 88.1 | 54.39 | |
| 10 f. + LR fadingTarget Frames=10, Learning Rate=fading2025.12 | 87.53 | 53.82 | |
| Target domainAdaptation Protocol=Target domain baseline2025.12 | 84.66 | — | |
| Diverse Frame Selection AlgorithmNumber of frames=100, Learning rate strategy=constant2025.12 | 82.23 | 52.99 | |
| Diverse Frame Selection AlgorithmNumber of frames=37, Learning rate strategy=constant2025.12 | 81.39 | 52.15 | |
| SEE2025.12 | 79.39 | 45.68 | |
| Diverse Frame Selection AlgorithmNumber of frames=10, Learning rate strategy=constant2025.12 | 79.22 | 49.98 | |
| Bi3DNumber of frames=2812025.12 | 78.03 | 48.79 | |
| CMT2025.12 | 74.53 | 45.29 | |
| ST3D2025.12 | 64.78 | 35.54 | |
| Stat-Norm2025.12 | 63.6 | 34.36 | |
| Source onlyadaptation_type=Baseline2025.12 | 33.71 | — | |
| Source onlyAdaptation Protocol=Source only baseline2025.12 | 29.24 | — |