🚀 AI-powered player tracking and automated statistics for futsal
PlayerTracker is an innovative platform that uses artificial intelligence to automatically detect players and generate advanced statistics from futsal match videos.
Designed for data analysts, coaches, and clubs, PlayerTracker provides an automated and intuitive way to track team performances, offering valuable insights without requiring manual intervention.
This project was developed for commercial purposes and was also used as a final-year academic project in software engineering.
- Player detection from match videos 🎥
- Automatic generation of game statistics 📊
- Identification of key events (goals, passes, shots, possession) ⚽
- Recent matches analysis at a glance
- Performance comparison across multiple matches
- Dynamic visual analytics, including:
- Goal ratio per match
- Possession rate
- Heatmaps for player and ball movements 🔥
- Upload matches (video files must meet specific requirements)
- Edit match details (team names, date, etc.)
- Export match reports as PDFs 📄
- Add and manage team players & coaching staff 👥
- Team Management page to set up club details
- Secure authentication system 🔑
- Role-based access (analyst, admin)
- Multi-user support (add staff members to a team)
PlayerTracker leverages a modern full-stack architecture, combining a powerful backend, an intuitive frontend, and AI-driven analytics.
- Adonisjs – API development
- MinIO – Object storage for video processing
- PostgreSQL - Data storing
- React – React-based frontend with client-side rendering
- TailwindCSS – Clean and responsive UI design
- YOLOv11 (You Only Look Once) – Player & ball detection
- OpenCV – Video & image processing
- Roboflow – AI model labeling & optimization
- Docker - Containerization
- Frontend: PlayerTracker-Frontend
- Backend: PlayerTracker-Backend
- AI Processing: AI-tracker
Ensure you have the following installed:
- Node.js (v16+) and npm (or yarn)
- Docker & Docker Compose (recommended)
- MinIO (for video storage)
- FFmpeg (for video processing)
Users sign up and log into their secure dashboard.
- Go to the Uploader page
- Enter match details (teams, date, etc.)
- Select a video file and click Upload
- AI automatically analyzes the match 🧠
- View all generated statistics on the Statistics page 📊
- Compare performance across matches
- Explore dynamic charts & heatmaps 🔥
- Match reports can be exported as PDFs 📄
- Add/edit players & coaching staff 👥
- Customize club details 📋
✅ MVP available with core features 🎉
🔜 Upcoming Enhancements:
➜ Pass tracking 🎯
➜ Advanced analytics on shots, goals & duels ⚽
➜ Player ID tracking for individual performance monitoring 🔍
➜ Enhanced AI models for improved accuracy 🤖
We are open to:
✅ Collaborations with futsal clubs & federations
✅ Investment opportunities to expand & enhance the platform
✅ Acquisition of the MVP and concept under specific conditions
📩 If you’re interested, feel free to reach out to us!
- The full source code remains private within our GitHub organization.
- No API keys or sensitive data are shared in this repository.
- LICENSE.md ensures intellectual property protection.
📢 If we decide to release an open-source version, only selected non-critical components will be public.
- Nawfel Ajari – Project Manager
- Kristian Vasiaj – Fullstack, AI Specialist
- Ismael Bouzrouti – Backend, Devops & AI Specialist
- Mehdi Merkachi – Fullstack & Performance Optimization
- Soufiane Hamoumi – Frontend & UX/UI
The analysis phase of PlayerTracker began in October 2024. The contributors worked on this project across five different sprints between February and March 2025, following the Scrum methodology combined with Kanban for task management and workflow optimization.
A huge thank you to everyone contributing to PlayerTracker! 🚀
📧 Email: info@nainnovations.be
🔗 GitHub: n4wf3l