Comparative Assessment of MediaPipe and Open-Source Tools for Biomechanical Analysis Based on Computer Vision (#2036)

Authors

Aguilar López, Karla Sarahi

Lainez Soto, Celeste Odeth

Vallecillo Acosta, Juan Fernando

Abstract

Understanding human movement is essential in the study of biomechanics, particularly in academic settings, where students and researchers seek effective methods to analyze human motion. This project explores the use of open-source tools to study lower-limb biomechanics, alongside the development of a custom program based on MediaPipe as a computer vision tool. The goal is to create an accessible Python application to analyze leg movement, focusing on knee flexion and extension during squats and gait assessment. Through real-time computer vision, this work aims to provide universities with a simple and cost-effective alternative for integrating motion analysis into teaching and research. Rather than achieving absolute accuracy, the focus is on understanding how MediaPipe can complement other open-source tools and its role in biomechanics education. However, while its accesibility and ease of use make it a promising option, further research and experimentation are necessary to explore its full potential and possible improvements for human movement analysis.