Abstract
Objective: This study examines the use of artificial intelligence-supported performance analysis systems in amateur and professional athletes, comparing existing technologies and evaluating trends in the literature within a comprehensive framework. The increasing use of artificial intelligence-based tools in sports science enables the objectification of training processes and the multidimensional measurement of performance. Therefore, this study details the contributions of artificial intelligence-based analysis systems to sports and the limitations encountered. Scope: The research covers a broad literature spanning multiple sports disciplines (soccer, basketball, swimming, athletics, racket sports, combat sports, etc.) through academic articles, technical reports, and experimental studies published between 2015 and 2025. The studies examined focus on areas such as motion analysis, injury risk prediction, load tracking, biomechanical measurements, computer vision-based performance evaluation, and machine learning-based decision support processes. Method: The literature review was conducted using a systematic approach. Databases such as IEEE Xplore, PubMed, Scopus, and Google Scholar were searched using the keywords "AI-based performance analysis," "machine learning in sports," and "computer vision athlete monitoring." Studies containing experimental findings, clearly defining artificial intelligence methods, and aiming to measure athlete performance were selected as inclusion criteria. Conclusion: The data obtained after the literature review shows that AI-supported performance analysis systems can measure with high accuracy, especially in professional sports. It reveals that they particularly improve decision-making processes and accelerate technical development. Computer vision methods stand out in motion analysis and technical error detection, while machine learning models excel in performance prediction, load analysis, and injury risk prediction. Low data quality in amateur athletes, equipment costs, and problems accessing expert support were identified as significant limitations. Overall, AI systems show strong potential for increasing training efficiency, providing personalized feedback, and reducing injury risk. However, ethical responsibilities, data privacy, and the lack of low-cost, accessible models are key research areas to be addressed in the future.

This publication has 23 references indexed in Scilit: