version: 0.1.0
Football Player Detection (YOLO11m)
This model is a fine-tuned YOLO11m object detector trained to identify key entities in football (soccer) match images.
It was trained on the martinjolif/football-player-detection dataset and is suitable for sports analytics tasks such as player tracking or match analysis for example.
Detected Classes
- player
- goalkeeper
- referee
- ball
Training Details
- Base model: Ultralytics YOLO11m
- Task: Object Detection
- Dataset: martinjolif/football-player-detection
- License: AGPL-3.0
Evaluation Results
Results on the test split using standard object detection metrics (Precision, Recall, F1, mAP50, mAP50-95):
| Class | Images | Instances | Precision | Recall | F1-score | mAP50 | mAP50-95 |
|---|---|---|---|---|---|---|---|
| ball | 23 | 24 | 0.8912 | 0.6833 | 0.7735 | 0.6799 | 0.3380 |
| goalkeeper | 19 | 19 | 0.9473 | 0.9463 | 0.9468 | 0.9413 | 0.8024 |
| player | 25 | 500 | 0.9576 | 0.9940 | 0.9755 | 0.9937 | 0.8737 |
| referee | 25 | 56 | 0.9059 | 0.9643 | 0.9342 | 0.9888 | 0.7741 |
Prediction Examples:
Notes
- Excellent performance on player detection, with very high recall and mAP.
- Ball detection is more challenging due to small object size and motion blur.
Intended Use
- Football match analytics
- Player, referee, and ball tracking
- Annotation assistance for sports datasets
- Research in sports computer vision
Limitations
- Performance may drop on non-broadcast footage, unusual camera angles, or low-resolution images.
- Trained specifically for football (soccer); not intended for other sports without further fine-tuning.
Usage
wget https://huggingface.co/martinjolif/yolo-football-player-detection/blob/main/yolo-football-player-detection.pt
from ultralytics import YOLO
model = YOLO("yolo-football-player-detection.pt")
Citation
If you use this model or the dataset, please cite the dataset authors and Ultralytics YOLO accordingly.