Object Detection on KITTI Snow (test)
89.5AP (Car)ActMAD
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ActMADBackbone=YOLOv3, Pre-trained=KITTI-Clear, Adaptation=Active Model Adaptation, Learning Rate=1e-4, Batch Size=302022.11 | 89.5 | 78 | 82.6 | 57.8 | 38 | 53.4 | 58.8 | 58.1 | 64.5 | |
| TTTBackbone=YOLOv3, Pre-trained=KITTI-Clear, Adaptation=Self-supervised Rotation Prediction2022.11 | 77.2 | 53.2 | 60.6 | 48.4 | 29.7 | 37.1 | 43.2 | 31.1 | 47.5 | |
| NORMBackbone=YOLOv3, Pre-trained=KITTI-Clear, Adaptation=Normalization2022.11 | 75.5 | 51 | 51.7 | 46.8 | 21.7 | 34.9 | 43.4 | 34.3 | 44.9 | |
| DUABackbone=YOLOv3, Pre-trained=KITTI-Clear, Adaptation=Dynamic Uncertainty Adaptation2022.11 | 64.6 | 38.9 | 49.3 | 44 | 20.8 | 22.8 | 27.8 | 25.4 | 36.7 | |
| SourceBackbone=YOLOv3, Pre-trained=KITTI-Clear, Adaptation=None2022.11 | 54.8 | 27.8 | 31.7 | 35.7 | 1.3 | 15.7 | 18.2 | 13.3 | 24.8 |