Abstract
Cricket is a type of sport that can be described as a bat and ball game played by two teams, where the objective of the game is to accumulate as many runs possible. Runs are accumulated through a batsman’s ability to perform diff t cricketing strokes. The strokes being performed correctly may potentially lead to the ac- cumulation of more runs, which will increase a team’s chance of winning. The increased viewership within the domain of cricket has created pressures for players to perform, which has resulted in incidents of cheating and doping to gain a competitive advantage. Additionally, certain areas, such as the Caribbean (West Indies), lack dedicated research and resources in the area of automated anal- ysis. Thus, these areas are at an unfair disadvantage over other regions, where sufficient research and resources are available, such as Australia and England. Areas that lack sufficient research and resources have yet to discover an effective means for automated analysis of cricket strokes. Therefore, there is a lack of research and resources that determines which algorithms are an effective means for automated analysis of cricket strokes. This research study aims to create a cricket stroke recognition model using computer vision methods. The research aim can be achieved by analyzing the literature within the research domain. The literature helps identify the problem areas, the different types of cricketing strokes relevant to the study, which includes the block, cut, drive, glance, and hook/pull strokes, and investigates the dif- ferent computer vision methods for achieving stroke recognition. The development of a prototype is used to determine the feasibility of the proposed model and evaluate the different approaches for achieving stroke recognition...
M.Sc. (Information Technology)