Tennis Video Analysis System With YOLO, PyTorch, and CNN

🎾 Excited to share my latest project - Tennis Video Analysis System With YOLO, PyTorch, and CNN! 🚀

I've developed an innovative solution that analyzes tennis players in videos to measure their speed, ball shot speed, and number of shots. Using YOLO from Ultralytics and Roboflow for player and tennis ball detection, along with a CNN developed by PyTorch for court keypoint extraction, Pandas for calculating statistics of player movement speed, ball shot speed, and number of shots, and OpenCV for presenting those statistics on the video, this project delivers accurate insights into player performance.

This project goes beyond traditional video analysis. By leveraging state-of-the-art technologies like YOLO, PyTorch, and CNNs, it provides detailed insights into tennis player performance, helping players and coaches make data-driven decisions to improve their game.

Witness the power of this system with real-time statistics overlaid on the video, offering a comprehensive view of player dynamics and shot analysis

YOLO from Ultralytics and Roboflow for precise player and tennis ball detection

PyTorch CNN for robust tennis court keypoint extraction

Pandas for calculating and analyzing player movement speed, ball shot speed, and shot count

OpenCV for presenting statistical insights on the video

With detailed statistics on player movements, shot speed, and frequency, this system empowers players and coaches to identify strengths and weaknesses, refine strategies, and enhance overall performance.

Tennis ball detector with YOLO: Research_Notebooks/Ball_Detector.ipynb

Tennis court keypoint with Pytorch: Research_Notebooks/Tennis_Court_Points_Detector.ipynb

Python 3.8

Ultralytics

PyTorch

Pandas

NumPy

OpenCV