DOI: 10.56639/jsar.1994860 ISSN: 2687-6027

Artificial Intelligence Applications in Sports Refereeing: A Systematic Review and Methodological Comparison

Meltem Tarım Karakuzu, Alparslan Erman
Show PDF Cite
Artificial Intelligence (AI) technologies are increasingly being used in sports refereeing to support decision-making, object and event tracking, and player- or team-related analysis. This systematic review aimed to examine the application areas of AI technol-ogies in sports refereeing across different sports, identify the analytical methods em-ployed, and compare their methodological performance based on reported accuracy rates. The literature search and study selection process were conducted in accordance with PRISMA guidelines. Scopus, Web of Science, IEEE Xplore, EBSCO, and Google Scholar databases were searched using the string “((Artificial Intelligence) OR (machine learning) OR (deep learning)) AND Sport AND referee”. A total of 404 records were initially identified; after duplicate removal, 374 studies remained. Following title and abstract screening, 45 articles were assessed in full text, and 14 studies met the inclusion criteria. The findings showed that AI applications were predominantly implemented in football. The primary application areas were Referee Decision Support Systems (7 stud-ies), Object and Event Tracking (5 studies), and Player and Team Analysis (2 studies). The most frequently used methods included CNN, LSTM, and YOLO-based deep learning models. Image-processing-based models, particularly YOLO-based approaches, gener-ally demonstrated higher accuracy compared to sensor-based systems. Overall, AI has potential to enhance the accuracy of refereeing decisions and operational efficiency; however, performance limitations, data quality issues, and ethical concerns remain sig-nificant challenges.

More from our Archive