Project Abstract

Motivated by the opportunity to enhance sports coaching through technology, this project focuses on real-time feedback in boxing, using the Mediapipe framework for rapid pose detection. By merging machine learning with live video input, the system aims to revolutionize coaching experiences, providing insights previously available only through human observation. Positioned as a comparative study, this research evaluates the efficacy of the Mediapipe framework as well as some analytical techniques (including machine learning and algorithmic approaches) against traditional human coaching in boxing form assessment, addressing questions of accuracy, reliability, and usability. Employing a mixed-methods approach, it combines quantitative analysis of boxing metrics with qualitative feedback from coaches and participants. Initial findings suggest the developed system accurately analyses key form aspects, offering consistent assessments, particularly for novices, while human coaching provides nuanced feedback. This research contributes to the understanding of how technology can augment traditional coaching methods, potentially democratizing access to high-quality coaching and paving the way for future innovations in sports training and performance analysis.

Keywords: Computer Vision, Real-time Analysis, Machine Learning

 Conference Details

Session: Poster Session A at Poster Stand 31

Location: Sir Stanley Clarke Auditorium at Tuesday 7th 13:30 – 17:00

Markers: Daniele Cafolla, Eike Neumann

Course: BSc Computer Science, 3rd Year

Future Plans: I’m continuing studies