Felix Gomez Guillamon Jurado

Deep learning Pose-detection model for sports training assistance

Deployed a Real-Time Pose-Detection Deep Learning model with TensorFlow, Keras and OpenCV which is able to detect and render up to 17 different body keypoints.


  • Used TensorFlow metal for ARM-based mac devices.


  • The Model is able to detect up to 6 different people at the same time. With a total of 102 rendered keypoints.


  • This model will aid sports coaches to detect weaknesses in their athlete’s body technique.


  • This model is applicable to most technical-based sports: Tennis, Golf, Football, Swimming, etc.


  • Model is usable in either real-time video or by uploading any video.


  • Created associated ‘.yaml’ file for creating a working conda virtual environment.


  • Included other relevant documentation such as a ‘readme.md’ file.
football pose detection render
jupyter pose detection
felix gomez guillamon
felix gomez guillamon