Object Detection on SkyFusion-plane 2024 (test)
98.27AP50 (bbox)LEVIRDetNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LEVIRDetNetEvaluation Protocol=Direct-transfer2026.06 | 98.27 | 58.29 | 72.25 | |
| DEIMv2 (DINOv3 + ViT-Base)Backbone=ViT-Base, Pre-training=DINOv3, Evaluation Protocol=Target-supervised2026.06 | 97.58 | 64.62 | 78.79 | |
| YOLOv5xuEvaluation Protocol=Target-supervised2026.06 | 96.3 | — | — | |
| YOLOv12xEvaluation Protocol=Target-supervised2026.06 | 94.52 | 61.72 | 76.7 | |
| Deformable DETREvaluation Protocol=Target-supervised2026.06 | 94.4 | 54.11 | 70.53 | |
| DINOEvaluation Protocol=Target-supervised2026.06 | 94.28 | 59.67 | 72.75 | |
| DynamicVis-baseEvaluation Protocol=Target-supervised2026.06 | 92.8 | 69.7 | 75.4 | |
| DynamicVis-largeEvaluation Protocol=Target-supervised2026.06 | 92.7 | 71 | 75.8 | |
| RTMDetEvaluation Protocol=Target-supervised2026.06 | 92 | 60.7 | 70 | |
| YOLOXEvaluation Protocol=Target-supervised2026.06 | 91.47 | 56.4 | 67.04 | |
| DEIMv2 (DINOv3 + CNN)Backbone=CNN, Pre-training=DINOv3, Evaluation Protocol=Target-supervised2026.06 | 91.47 | 53.9 | 68.94 | |
| FCOSEvaluation Protocol=Target-supervised2026.06 | 91.36 | 54.64 | 66.87 | |
| EfficientDetEvaluation Protocol=Target-supervised2026.06 | 91.32 | 48.55 | 60.87 | |
| YOLOv3Evaluation Protocol=Target-supervised2026.06 | 91.28 | 49.13 | 65.27 | |
| DEIMEvaluation Protocol=Target-supervised2026.06 | 91.05 | 50.39 | 67.27 | |
| HrNetEvaluation Protocol=Target-supervised2026.06 | 90.64 | 59.15 | 69.47 | |
| Faster R-CNNEvaluation Protocol=Target-supervised, Reference=[30]2026.06 | 90.52 | 53.26 | 66.38 | |
| TridentNetEvaluation Protocol=Target-supervised2026.06 | 90.37 | 51.21 | 64.33 | |
| Fast R-CNNEvaluation Protocol=Target-supervised2026.06 | 90.1 | 45.98 | 61.59 | |
| SSDEvaluation Protocol=Target-supervised2026.06 | 89.66 | 46.56 | 62.23 | |
| YOLOFEvaluation Protocol=Target-supervised2026.06 | 83.38 | 26.47 | 54.3 | |
| RetinaNetEvaluation Protocol=Target-supervised2026.06 | 66.48 | 13.26 | 35.98 | |
| Sparse R-CNNEvaluation Protocol=Target-supervised2026.06 | 51.27 | 20.81 | 28.28 |