Bryson DeChambeau Trains His Swing With 3D Skeletal Tracking
Google Cloud built a system that runs deep learning models on 2D and 3D video. It marks 30 body, club, and ball points at the top of the backswing, at impact, and through the finish. Shot outcome data is logged automatically for each repetition.
Athletes stop guessing about form. They replace vague video review with labeled joint angles and ball flight numbers. Training decisions shift from feel to measurable deltas.
Bryson DeChambeau and the Google Cloud sports team tested the pipeline on range sessions. The output gave him frame-by-frame joint coordinates and club-path metrics for every swing.
Step 1: Record a swing on any smartphone and upload the clip to Google Cloud Video AI at cloud.google.com/video-intelligence. Step 2: Enable the Pose Detection feature so the model returns 30 key points per frame. Step 3: Export the JSON and open it in a spreadsheet to compare joint angles across ten swings.