Table 1 Model training results at a confidence threshold of 0.320
From: Using spatial video and deep learning for automated mapping of ground-level context in relief camps
Metric | Value |
|---|---|
F1-Score | 0.82 |
Precision | 0.84 |
Recall | 0.79 |
mAP@0.5 | 0.84 |
mAP@0.5:0.95 | 0.45 |