Multi-UAV Collaborative Panoptic Segmentation on Air-Co-Pred (test)
65.52IoU (Short)MaskDINO (SwinL)
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MaskDINO (SwinL)Backbone=Swin-L, Number of UAVs=4, Fusion Strategy=BEV fusion, Top-k Ratio=25%2026.03 | 65.52 | 57.66 | 61.59 | 54.1 | 50.63 | 52.37 | 6.14 | |
| MaskDINO (ResNet50)Backbone=ResNet-50, Number of UAVs=4, Fusion Strategy=BEV fusion, Top-k Ratio=25%2026.03 | 63.99 | 55.97 | 59.98 | 54.24 | 48.23 | 51.24 | 6.14 | |
| EfficientNet (4 UAVs)Backbone=EfficientNet, Number of UAVs=4, Fusion Strategy=BEV fusion, Top-k Ratio=25%2026.03 | 60.58 | 53.14 | 56.86 | 51.8 | 45.89 | 48.85 | 6.14 | |
| EfficientNet (2 UAVs)Backbone=EfficientNet, Number of UAVs=2, Fusion Strategy=BEV fusion, Top-k Ratio=25%2026.03 | 50.59 | 40.82 | 45.71 | 43.43 | 35.37 | 39.4 | 2.05 | |
| MaskDINO (SwinL) w/o FusionBackbone=Swin-L, Number of UAVs=1, Fusion Strategy=w/o Fusion2026.03 | 45.57 | 34.3 | 39.94 | 39.5 | 30.26 | 34.88 | 0 | |
| EfficientNet w/o FusionBackbone=EfficientNet, Number of UAVs=1, Fusion Strategy=w/o Fusion2026.03 | 41.31 | 33.11 | 37.21 | 35.07 | 28.29 | 31.68 | 0 |