Detecting a player once is one thing — following that same player continuously, for ninety minutes, is a much harder problem. SPOIT's tracking engine takes the output of object detection and turns it into continuous motion data, frame by frame, for the entire match.
Continuous, Not Sampled — SPOIT tracks continuously rather than sampling positions every few seconds, meaning the movement data reflects what actually happened.
Handling a Chaotic, Moving Scene — With 22 players constantly changing position and occluding each other, SPOIT's tracking engine is built to hold onto player identity even in crowded, unpredictable moments.
Conclusion: Real-time tracking is what turns a video into a dataset — the basis for team stats, individual player data, and 2D tactical maps.
Made with Bullet