Bryan Zhao

Pose Estimation for Sprinter Form Comparisons

Started May 29, 2024

Abstract or project description

How can we use pose estimation to evaluate sprint form for an athlete's top sprint position? That was the question I hoped to answer with my project. As a high-school track athlete specializing in sprinting, I spend hours each week improving my running technique and increasing my sprinting power with the help of my coaches and peers. However, with my knowledge of data science and machine learning, I also wanted to contribute to growing discoveries in sprinting, so I started this project with Polygence to better understand the fundamentals of sprinting. With my mentor Atharv, we collected footage from high-school athletes and Olympic-level athletes to cross-analyze what differentiated an amateur from a professional. By using pose estimation to analyze the angles exhibited by both variations of athletes and collecting sprint-time data of high-school athletes, we hoped to draw a consensus on what made a sprinter faster.