Well, Actually, Your Golf Swing Needs 30 Tracking Points and a Cloud Computing Budget
Bryson DeChambeau partnered with Google Cloud to deploy an AI system employing deep learning and proprietary 2D/3D biomechanical models. The system tracks over 30 critical points across the golfer's body, club, and ball to analyze swing phases including top of swing, impact, and follow-through. Source: https://cloud.google.com/transform/the-ai-crusher-bryson-dechambeau-on-his-experiments-with-ai-and-the-future-of-sports-performance
This demonstrates the principle of multi-modal biomechanical analysis: combining computer vision with physics-based modeling to break continuous motion into discrete, optimizable phases. For your workflow, consider how any complex process can be decomposed into trackable segments rather than treated as a monolithic black box.
Bryson DeChambeau, professional golfer, in collaboration with Google Cloud. The system uses their proprietary models; no specific performance metrics or tournament results from this partnership are detailed in the source.
Step 1: Open your phone's camera app and record a 10-second video of yourself performing any repetitive motion (a golf swing, tennis serve, or even a walking stride). Step 2: Download a free pose estimation app such as PoseNet or use a free trial of a sports analysis app like Hudl Technique; import your video and observe which body points the software automatically detects. Step 3: Manually mark three distinct phases of your motion (preparation, execution, follow-through) and compare the software's joint angle measurements across these phases to identify where your form deviates most from a reference video you find online.