Classification of Cricket Shots using Convolutional Neural Network

Abstract

Cricket is a bat-and-ball game played between two teams of eleven players each on a circular or oval shaped field with a rectangular 22-yard pitch at the center. Originating in England in the 16th century, it has been evolving into a globally popular sport till now, particularly in countries like India, Australia, and England. The game comprises formats like Test cricket, One Day Internationals (ODIs), and Twenty20 (T20) cricket, each varying in duration and style. The objective is the batting team to score runs while the bowling team aims to dismiss the batsmen. The Key elements include batting, bowling and fielding, with matches being officiated by umpires. This research presents a comprehensive overview of detecting cricket shots using InceptionV3 for the classification of images. After trying MobileNet, VGG19, ResNet, DenseNet and InceptionV3 layer. Only MobileNet turned out to be better in our dataset. This Research can be significant in the advancements of AI Technology in Robotics and can be helpful to the coaches.