Full Custom YOLO Detection Model
Model Description
This model is a custom-trained YOLO object detection model for multi-class detection tasks on a specialized dataset.
It is trained for fine-grained object detection using bounding box annotations across multiple visual anatomy classes.
Intended Use
- Object detection on custom anatomical datasets
- Bounding box classification for fine-grained region detection
Limitations
- Trained on a custom manually annotated dataset across 7 classes
- Performance may degrade on unseen domains or different distributions
- Evaluation is based on a controlled dataset of ~50 images
Evaluation Results
Overall Metrics
- Precision: 0.987
- Recall: 0.999
- mAP@50: 0.995
- mAP@50-95: 0.794
Per-Class Results
| Class | Images | Instances | Precision | Recall | mAP50 | mAP50-95 |
|---|
| all | 50 | 253 | 0.987 | 0.999 | 0.995 | 0.794 |
| face | 41 | 49 | 0.989 | 1.000 | 0.995 | 0.862 |
| nipple | 33 | 64 | 0.985 | 0.995 | 0.995 | 0.785 |
| mouth | 41 | 50 | 0.971 | 1.000 | 0.995 | 0.793 |
| eyes | 41 | 49 | 0.990 | 1.000 | 0.995 | 0.832 |
| navel | 28 | 30 | 0.989 | 1.000 | 0.995 | 0.697 |
| anus | 11 | 11 | 0.995 | 1.000 | 0.995 | 0.792 |
