Bryson DeChambeau Uses Deep Learning to Track 30 Body Points in Real Time
Bryson DeChambeau's team built a system with Google Cloud deep learning models. The models analyze 2D and 3D video frames and track over 30 anatomical landmarks on the golfer, club, and ball. They record swing phases such as top of backswing, impact, and follow-through.
Athletes and coaches now treat motion capture as a repeatable data product instead of a one-off video review. The workflow shifts from subjective feel to measurable joint angles and timing metrics that can be compared across sessions.
Bryson DeChambeau and the Google Cloud sports performance team have published swing analytics that show measurable improvements in club path consistency during testing sessions.
Step 1: Record a swing with any smartphone camera at 60 fps or higher. Step 2: Upload the video to Google Cloud Video AI or MediaPipe Pose and run the pose detection model. Step 3: Export the 30-point landmark CSV and compare joint angles between two sessions using a spreadsheet. https://cloud.google.com/video-intelligence