Automating cricket scorecards with deep neural network: a next-generation approach

This work presents an approach to the automation of scorecards in cricket using CNN.The objective will be developing such a system that will extract information that is relevant automatically and efficiently from actual cricket matches and provide the creation of a scorecard. A CNN model proposed in this study will be trained on a cricket match data set of considerable size, which will help to recognize and analyze different features of the game, which may consist of runs scored, wickets taken, and player performance. Automation in this area, in combination with this technique, could lead to a reduction in the amount of time and effort that cricket statisticians and broadcasters have to put in, while ensuring accurate, up-to-date scorecards with an accuracy of 95%, which is higher than currently available 7%.