ECE23: BARPATH

Project Description:

Olympic weightlifting requires precise kinematic analysis to optimize performance and prevent injury, yet traditional motion capture systems demand expensive equipment and controlled environments. Barpath is a computer vision system that uses a custom-trained YOLO model to detect a barbell, with Mediapipe pose estimation to capture an athlete's movements. This data is then run through an analysis system trained on elite-level weightlifters, allowing novice and intermediate level athletes to accessibly compare their technique to these elite references. The system can automatically detect which of the three expected movements are being performed, and, using multiple classifiers across the data gathered, will return various errors or deviation observed between the user and their selected model athlete, with actionable steps to rectify the deviation.

Advisor/Instructor:

Sponsor:

Team Members:

Poster:

Table #: