Model Analysis Report
- Generated at:
2026-08-31T05:10:37.559752
- Experiment dir:
runs/rotated_faster_rcnn/20260831-035851
- Checkpoint:
runs/rotated_faster_rcnn/20260831-035851/checkpoints/best_mAP_0.89.pth
- Checkpoint modified:
2026-08-31T04:57:34.306132
- Config:
runs/rotated_faster_rcnn/20260831-035851/config.json
Source data
- Data root:
/path/to/data/HRSC2016
- Data split:
val
- Total images:
453
- Total ground truth objects:
1228
- Total predictions:
1545
Evaluation setup
- mAP / PR matching IoU (rotated boxes, VOC-style; not NMS IoU):
0.50
- NMS IoU (deduplication):
0.10
- Threshold sweep:
0.0 to 1.0 step 0.05
Key outcomes
- Best threshold (F1):
0.9000
- Precision at best threshold:
0.9311
- Recall at best threshold:
0.9023
- F1 at best threshold:
0.9165
- F2 at best threshold:
0.9079
- mAP50:
0.8877 (88.77%)
GT alignment (mean best IoU vs raw detections)
- Global mean best IoU (any class):
0.7848
- Global mean best IoU (same class):
0.7848 (median 0.8183)
Per-class breakdown (each GT: max rotated IoU vs detections on the same image):
| Class |
gts |
mean_any |
mean_same |
med_same |
ship |
1228 |
0.7848 |
0.7848 |
0.8183 |
| global |
1228 |
0.7848 |
0.7848 |
0.8183 |
Per-class metrics (mAP50)
| Class |
gts |
dets |
recall |
AP |
ship |
1188 |
1545 |
0.962 |
0.8877 |
| mAP |
|
|
|
0.8877 |
Per-class best thresholds (max F1 over the same sweep)
| Class |
Threshold |
Precision |
Recall |
F1 |
TP |
FP |
FN |
ship |
0.9000 |
0.9311 |
0.9023 |
0.9165 |
1108 |
82 |
120 |
Confusion matrix
Computed at score threshold 0.9000 and IoU 0.50.
Rows are ground-truth classes; columns are predicted classes. The False Positive row contains unmatched detections; the Missed column contains unmatched GTs.
| Actual \ Predicted |
ship |
Missed |
ship |
1108 |
120 |
False Positive |
82 |
0 |
Artifacts
- Predictions JSON:
predictions.json
- Analysis JSON:
analysis_iou0.50.json
- PR curve:
pr_curve.png
- Threshold metrics:
threshold_metrics.png
Notes
- Global threshold selected by maximizing F1; tie-breaks favor recall, then lower threshold.
- Per-class table: best threshold per class maximizes F1 on the same threshold grid (see
best_threshold_per_class in the analysis JSON).
- Precision/recall are computed using class-aware IoU matching with one-to-one assignment.