Object Detection on Golden 443 train samples (test)
89.1mAP@0.50YOLO
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| YOLOArchitecture=YOLO, Training Strategy=Golden + ECP (Trained on Golden dataset with ECP hyperparameterization)2026.05 | 89.1 | 88.4 | 81.9 | 85.1 | 50.5 | |
| SatlasNetArchitecture=SatlasNet, Training Strategy=Trained directly on Golden dataset2026.05 | 87.6 | 80.9 | 90.5 | 85.4 | 59.2 | |
| Weakly Supervised School Detection Framework (YOLO)Architecture=YOLO, Training Strategy=Ours (Trained on auto-labeled dataset and fine-tuned on Golden set)2026.05 | 87.6 | 89.3 | 79.3 | 84 | 66.2 | |
| Faster R-CNNArchitecture=Faster R-CNN, Training Strategy=Golden + ECP (Trained on Golden dataset with ECP hyperparameterization)2026.05 | 78.6 | 56.9 | 83.3 | 67.6 | 52.6 | |
| Weakly Supervised School Detection Framework (Faster R-CNN)Architecture=Faster R-CNN, Training Strategy=Ours (Trained on auto-labeled dataset and fine-tuned on Golden set)2026.05 | 78 | 64.2 | 81 | 71.6 | 60.2 | |
| Faster R-CNNArchitecture=Faster R-CNN, Training Strategy=Auto (Trained on auto-labeled dataset)2026.05 | 60 | 19.3 | 86.9 | 31.6 | 39.5 | |
| YOLOArchitecture=YOLO, Training Strategy=Auto (Trained on auto-labeled dataset)2026.05 | 46.4 | 47.8 | 51.2 | 49.4 | 33.2 |