Welcome to Football Vision AI! This project is all about bringing the excitement of football to the world of AI. Our system uses cutting-edge computer vision techniques to detect players, referees, goalkeepers, and the ball, providing real-time visualizations and insights into the state of the game. Whether you're a football enthusiast, a data scientist, or just curious about AI, this project has something for you!

• Player and Referee Detection: Our AI can accurately identify and track all players, referees and the ball on the field using YOLOv8n and BoTSORT.

• Goalkeeper Identification: Distinguish goalkeepers from other players for specialized analysis.

• Two teams Identification: Differentiate the two teams for targeted analysis using PCA and KMeans only.

• Real-time Visualizations: Generate dynamic visual representations of the match state, making it easy to understand the flow of the game.

This project is an implementation based on a YouTube tutorial with some modifications and enhancements. You can watch the original tutorial here. Below are the key differences:

• Improved player and referee detection speed while maintaining accuracy by utilizing YOLOv8n instead of YOLOv8x.
• Enhanced keypoint detection efficiency without compromising accuracy by employing YOLOv8n Pose instead of YOLOv8x Pose.
• Improved players clustering algorithm by using PCA and K-Means on the cropped players without ReID.
• Optimized real-time visualizations.

Many visualization were taken from Supervision Repo.

Tracker used BoTSort

• Python 3.7+
• OpenCV
• NumPy
• scikit-learn
• ONNXRuntime

1. Clone the repository:

   git clone  https://github.com/Supernova1744/Football_Vision_AI.git

2. Navigate to the project directory:

   cd Football_Vision_AI

3. Install the required dependencies:

   pip install -r requirements.txt

1. Prepare your video files.
2. Run the main script:

   python main.py --video videos\match1.mp4

3. View the results in the output directory.

Check out a demo video of our system in action! Place your video oath in the --video flag for the main script to visualize the analysis.

Watch Demo Video

If you have any questions or suggestions, feel free to open an issue or contact us at ali.samir.1744@gmail.com.