Analysis of running form with keypoint R-CNN

Abstract

Good running form is extremely important for runners to avoid injury and to improve performance. However, most casual runners do not have access to a running coach who can help to correct their running form. Hence, this project aims to use computer vision to analyze running form by using Pytorch's Keypoint R-CNN, a Human Pose Estimation model, which detects certain important points in the human body. This project creates a simple tool that casual runners can use to analyze their own running form.