Wicket Keeping Technique Analysis with Deep Neural Networks and Pose Estimations

R.M.V Ishara and Sumanathilaka T.G.D.K.

Classification, Machine learning, Technique analysis , Wicket-keeper, Convolution neural network, Deep learning.

18th February 2022

In this modern era, Artificial Intelligence has emerged as the next data analytics powerhouse. The use of Machine Learning and Computer Vision algorithms in data analytics have become a popular trend since their introduction. However, applying Deep Neural Nets to various sports data analysis tasks and studying their performance has yet to be investigated. When it comes to analysis, it’s usually a difficult and time-consuming process. Because of the sport’s growth with technology, Al-based wicket-keeping technique analysis has become a trending topic. This study is based on research into wicket-keeping technique analysis and the proposed use of Convolutional Neural Networks (CNN) and poses estimation models to present an alternative approach. As wicket-keeping is less researched in history, access to a proper dataset is challenging and his study attempts to create a proper dataset and proposed a novel architecture to predict the real-time pose estimation using Neural Network based approach. This technical analysis tool will help wicket-keepers of any level to identify their technique faults on their own and rectify them without the help of a physical coach and improve their technique to become better wicket-keepers. The proposed tool will also aid coaches of any level to identify technique faults of their wicket keepers and to improve their coaching strategies to make them better players.