For Leagues, Broadcasters, and Tech Founders. We build custom Computer Vision Engines that track the puck at 100mph, analyze player positioning, and automate advanced stats (Corsi/Fenwick) in real-time—no sensors required.
Hockey is the hardest sport to track for a reason. Off-the-shelf cloud APIs fail miserably due to three unique challenges:
A puck moves at 100mph and is often blurred or hidden by spray. Generic models can't see it. We train custom models specifically for Small Object Detection in high-contrast environments.
Players pile up in the crease or battle along the boards, blocking the view. Our logic uses Temporal Consistency to predict where the player/puck is, even when hidden for a split second.
Rink glass creates glare that confuses standard cameras. Our preprocessing algorithms filter out reflection noise to maintain a clean track.
From "Rink to Cloud" for startups, to "Broadcast Overlays" for media rights holders.
We turn video into structured XY data. Track Zone Entries, Face-off wins, and Shot Quality (xG) automatically without requiring expensive RFID chips in the jerseys.
Hockey requires high frame rates. We ingest 60fps or 120fps video to reduce motion blur, ensuring the puck appears as a distinct object rather than a smear.
We map the camera angle to a 2D model of a standard NHL/IIHF rink. This allows us to calculate real-world speed (mph) and distance from the blue line.
We run a specialized "Small Object Detection" model solely for the puck, separate from the player tracker. It uses trajectory prediction physics to "fill in the gaps" when the puck is blocked by a stick.
Our color-segmentation algorithms automatically separate Home (White) vs. Away (Dark) jerseys and link possession time to specific teams.
Fill the form below or Contact us at +1 408 365-4638 / email us via contact@folio3.ai
of Experience In the AI Domain
Delivered Worldwide
Client Satisfaction
Founded
Response Guaranteed
+1 408 365-4638
contact@folio3.ai
6701 Koll Center Parkway, #250 Pleasanton, CA 94566
