-
-
-
-
The gateway to Anact Ortho, providing one-click access to Live, Film Room, HooperIQ, and Recruit experiences.
-
Upload or stream a basketball game to receive automated officiating, real-time commentary, live scoring, and AI-powered analytics.
-
Upload or stream a basketball game, calibrate the court, and let AI automate officiating, scoring, and live commentary.
-
Study real game footage, test your basketball IQ, and receive AI-assisted feedback through interactive film analysis.
-
Upload game footage to generate AI-powered scouting reports, player analytics, highlights, and shareable performance cards.
Inspiration
Every weekend, millions of basketball games are played without referees, performance analytics, or accessible coaching tools. Pickup games end in arguments over out-of-bounds calls, youth leagues struggle to recruit officials, and advanced scouting systems remain locked behind expensive cameras, proprietary hardware, and enterprise-level budgets.
We asked a simple question:
What if a single courtside camera could replace an entire officiating crew, commentator, and scouting team?
That idea became Anact Ortho—an AI-powered sports intelligence platform that makes professional-grade game analysis accessible to everyone.
What We Built
Anact Ortho is a computer vision-powered web application that transforms ordinary basketball footage into live game intelligence.
Using AI, the platform understands what's happening on the court in real time—tracking players, detecting game events, generating commentary, and producing professional scouting reports from a single video.
Our platform consists of four integrated experiences:
🏀 Live
Upload a courtside video or connect a live camera feed to receive:
- Automated officiating
- Real-time scorekeeping
- AI-generated play-by-play commentary
- Live ball tracking
- Instant whistle detection
🎥 Film Room
Replay professional basketball games through an interactive coaching interface with synchronized game events, reconstructed timelines, and tactical overlays.
🧠 HooperIQ
Improve basketball IQ by studying real possessions.
Users pause film, draw defensive or offensive reads, submit decisions, and receive AI-assisted feedback using a Glicko-inspired rating system.
📈 Recruit
Upload game footage and automatically generate:
- Player heatmaps
- Jump height
- Release velocity
- Game highlights
- Shareable scout cards
No wearable sensors.
No smart basketball.
No expensive tracking systems.
How We Built It
At the heart of Anact Ortho is an AI-driven computer vision pipeline that continuously understands the state of a basketball game.
Rather than simply detecting objects, our system combines multiple computer vision models into a unified game-state engine capable of interpreting events as they happen.
Our pipeline performs:
- Court calibration using four-point homography
- Player pose estimation
- Basketball tracking during motion and occlusion
- Boundary detection
- Automated event recognition
- Real-time commentary generation
- Player performance analytics
Our technology stack includes:
| Layer | Technology |
|---|---|
| Frontend | React, TypeScript, Vite |
| Backend | FastAPI, Python |
| Computer Vision | OpenCV, MediaPipe |
| Real-Time Communication | WebSockets |
| Database | SQLite |
The platform was designed as a modular architecture so every feature—officiating, commentary, and scouting—shares the same AI understanding of the game.
Challenges We Faced
Building Anact Ortho required solving both engineering and computer vision challenges under hackathon time constraints.
Real-Time Computer Vision
Tracking a fast-moving basketball is significantly more difficult than tracking players. Motion blur, player occlusion, lighting conditions, and camera movement required multiple detection strategies to maintain reliable tracking.
Browser + Local AI
Our web application runs in the browser, while advanced video processing relies on a local Python computer vision backend.
Designing a seamless experience between these environments required WebSockets, API health checks, and graceful offline fallbacks.
Performance
Processing full-resolution basketball footage introduced significant latency.
We optimized our demo pipeline around efficient preprocessing, shorter clips, and lightweight computer vision models to maintain a responsive experience during judging.
User Experience
Computer vision isn't perfect.
Rather than hiding uncertainty, we designed interfaces that clearly communicate prediction confidence and provide meaningful fallback states whenever AI confidence drops.
What We Learned
This project taught us that great AI products aren't just about model accuracy—they're about user experience.
We learned that:
- Fast feedback matters more than perfect accuracy.
- Clear AI confidence builds user trust.
- Modular architectures dramatically accelerate development.
- Computer vision becomes far more valuable when connected to real user workflows rather than isolated model outputs.
Most importantly, we learned that advanced sports technology doesn't have to be reserved for professional organizations.
Real-World Impact
Professional player tracking systems often cost tens of thousands of dollars, making advanced analytics inaccessible to schools, community leagues, and recreational athletes.
Anact Ortho changes that.
With nothing more than ordinary basketball footage, players and coaches gain access to automated officiating, live commentary, performance analytics, and scouting tools previously available only to elite programs.
Our vision is simple:
Bring professional sports intelligence to every basketball court.
What's Next
Our roadmap includes:
- Cloud-hosted AI processing
- Multiplayer live games
- Multi-camera tracking
- Improved player identification
- Team analytics dashboards
- Coach collaboration tools
- Mobile companion experience
- Expanded support for additional sports
Log in or sign up for Devpost to join the conversation.