Convolutional Neural Networks

This chapter introduces convolutional neural networks (CNNs) and describes how they can be used in the context of sports analytics. CNNs are suitable for end-to-end learning on images or similarly structured data. CNNs can efficiently learn features of images based on pixel values and, for example, extract suitable features for a classification task. In this context, the models benefit from parameter sharing in the convolutional layers and exhibit translation equivariance and invariance properties. CNNs are thus suited for learning features from positional data of team sports, provided that the data is put into an appropriate structure.