AI and computer-based methods in performance evaluation of sporting feats: an overview

• Published in: The Artificial intelligence review Vol. 27; no. 1; pp. 57 - 70
• Main Author: Owusu, Gilbert
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Nature B.V 01.01.2007
• Subjects: Athletes; Automation; Biomechanics; Coaches & managers; Computers; Educational materials; Multimedia; Performance evaluation
• ISSN: 0269-2821; 1573-7462
• DOI: 10.1007/s10462-008-9068-3

Performance evaluation is a complex process, usually involving the analyses of large amounts of possibly subjective information. The complexity increases when the performances of more than one athlete are being evaluated. For example a coach in charge of twenty divers should be able to remember the strengths and weaknesses of each athlete. Given these difficulties, it is therefore not surprising that a number of computer-based systems have been developed to speed-up and improve the quality of performance evaluation. Most of these systems are visually based such that individuals working on performance analysis first record the motion in question by electronic means and then input these images into a computer for quantification and subsequent analysis. There seems to be enormous potential for AI (i.e. Artificial Intelligence) technologies to make a significant contribution in the analysis phase. Indeed AI technologies have been applied to performance evaluation in recent years, though their applicability has been limited for a variety of reasons. The main factor has been a lack of characterisation of the domain of performance evaluation. This paper reviews selected research and applications of computational models and AI technologies in particular in performance evaluation of sporting feats for individual based events. [PUBLICATION ABSTRACT]