Computer vision for basketball
Pose analysis. Ball tracking. Broadcast automation.
Bridging the gap between the eye-test
and counting stats.
There's a common debate in basketball analytics between the eye-test and counting stats. I think computer vision on game footage bridges it: meaningful insights from the plays themselves, for coaching, scouting, and training, without someone having to watch every minute of every game.
Why Muse
Basketball is my most consistent and biggest passion, and my muse for upskilling technologically, practicing discipline, and learning about life.
The work
Turn players
into ballers.
I started Muse while at Boston University, with the goal of democratizing access to sophisticated basketball coaching and analysis. Follow Through looks at the individual: your shot mechanics compared with NBA players. EyeBall looks at the group: following the ball through game film to find passes, possessions, and eventually plays, and to broadcast amateur games without a camera operator.
Follow Through
Shot-form analysis
Record your shot and get immediate feedback: your mechanics compared frame by frame with an NBA player, a 0-99 similarity score, and the three things to work on first.
Explore
EyeBall
Ball tracking + broadcasting
Follows the ball through game film to pick out passes and possessions, and uses the same tracking to turn a fixed wide-angle camera into a broadcast that pans with the action, no operator needed.
ExploreGet in touch
Let's talk about the game.
If you work with a team, a league, or a training program and have problems with game film analysis or player development, I'd love to hear about them. I'm also looking for guidance on securing multi-angle footage from pro leagues to train better models.