Views
No views yet
| Model | Purpose | Size |
|---|---|---|
yolo26m.pt | Player/ball detection | ~50MB |
yolo26m-pose.pt | Pose estimation (optional) | ~50MB |
1# Clone the repository
2git clone https://github.com/YOUR_USERNAME/football-analysis.git
3cd football-analysis
4
5# Install dependencies
6pip install -r requirements.txt
7
8# Download YOLO models
9# Models are automatically downloaded on first run
10
11# Run analysis
12python main.py input_video.mp4 output_video.mp4config.py:1detection_confidence = 0.3 # Player detection threshold
2ball_confidence = 0.2 # Ball detection threshold
3device = "mps" # Auto-detected: mps/cuda/cpu| Device | FPS |
|---|---|
| Mac M4 (MPS) | 15-20 |
| NVIDIA RTX 3080 | 25-30 |
| CPU (i7) | 3-5 |
1@software{football_analysis,
2 title={Football Match Analysis with YOLO v26},
3 year={2026},
4 url={https://github.com/David-dsv/football-analysis}
5}