Developing an AI-Powered Referee System for Boxing: Integrating Sensor Data and Machine Learning for Punch Recognition
Ilomäki, Leevi (2024)
2024
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2024061823659
https://urn.fi/URN:NBN:fi:amk-2024061823659
Tiivistelmä
The goal of this project was to develop a video assistant referee system for boxing that leverages Artificial Intelligence to recognize landed punches through video feed, sound, and touch. This thesis presents research findings and related topics from the touch perspective, detailing the methods employed to capture boxing data and the subsequent processing required to create a machine learning model capable of accurately recognizing and counting landed punches in a boxing match. In the thesis, various methods for capturing landed punches are examined, and key data parameters to collect are discussed.
A study conducted by students at the Vienna University of Technology on the Smart Boxing Glove served as a guideline for this project. This study outlined the development of a machine learning-based system capable of classifying different strike techniques and target objects through sensor data.
In conclusion, the research and methods for data collection proved successful, though further development is necessary to achieve a fully operational system.
A study conducted by students at the Vienna University of Technology on the Smart Boxing Glove served as a guideline for this project. This study outlined the development of a machine learning-based system capable of classifying different strike techniques and target objects through sensor data.
In conclusion, the research and methods for data collection proved successful, though further development is necessary to achieve a fully operational system.