tennis_analysis

Why we included this project

This tennis video analysis project is a solid, hands-on way to understand computer vision beyond toy examples and into a real-world scenario. It runs YOLO to detect players and the ball, a fine-tuned CNN to locate court keypoints, then fuses those detections into player speed, ball shot speed, and shot counts. The training notebooks walk through fine-tuning each model, so you can follow the full pipeline from data to inference. Teams exploring sports analytics can borrow the modular structure as a realistic template for other games.