A-Eye: Automating the Role of the Third Umpire in the Game of Cricket

Cricket is a popular sport that involves two teams, say Team A and Team B. Each team comprises eleven players, along with two field umpires. Based on the outcome of a toss, one of the teams, say Team A, initially bats (by using a wooden structure called a bat) in order to score some runs. These runs are scored while Team B is balling (by using a special type of a ball). Then, the roles get reversed: Team A balls and Team B bats, while trying to overcome the score posted by Team A. If this happens, then Team B wins. Otherwise, Team A wins. The balling team aims to get the batsmen out, i.e., dismiss them from batting, in order to minimize the scoring. One of the ways in which this could happen is a run-out, i.e., the batsman fails to enter a particular playing area, before the bowling team dislodges three wooden stumps in that area. It is generally very difficult for the umpires to detect this scenario through the naked eye. Hence, it is typically referred to a third umpire, who makes the ´Run-Out/Not-Out´ decision through video technology. This process typically consumes 30 seconds (or more) which disrupts the fast pace of the game, e.g., in Twenty20 cricket (a type of cricket game). In this paper, we have proposed and implemented a novel technology for cricket, called A-Eye, which automates the role of the third umpire. By applying A-Eye to a set of autonomously-filmed run-out videos, we have shown that it is extremely efficient as compared to the third umpire, and almost as accurate. Also, it has the potential to minimize decision errors made by third umpires, and is able to estimate a rating for the field umpires. These results have convinced our local cricket council to employ A-Eye within a professional cricket tournament, which will be held soon.