Bryson DeChambeau Turns Golf Swing Data into Deep Learning Models
DeChambeau uses Google Cloud deep learning to track more than 30 body, club, and ball keypoints in both 2D and 3D. The system labels every phase of the swing from address through finish and records the resulting ball flight. All data feeds back into iterative model training.
Athletes can replace subjective feel with quantitative kinematic benchmarks. Workflows shift from video review to continuous model retraining loops that surface micro-adjustments in milliseconds.
Bryson DeChambeau and the Google Cloud sports performance team report measurable gains in swing repeatability after feeding 3D pose estimates back into practice routines.
Step 1: Open Google Cloud Vertex AI and upload swing footage to a new Pose Estimation dataset. Step 2: Train a custom 2D-3D keypoint model using the 30-point skeleton template. Step 3: Export inference results to BigQuery and plot joint angles across frames at https://console.cloud.google.com/vertex-ai.