Soccer Jersey Number Recognition Using Convolutional Neural Networks

In this paper, a deep convolutional neural network based approach to the problem of automatically recognizing jersey numbers from soccer videos is presented. It is meant as a tool for subsequent automatic player identification approaches that utilize jersey numbers together with knowledge about teams and the jersey numbers of their players. Two different jersey number vector encoding schemes are presented and compared to each other. The first treats every number as a separate class, while the second one treats each digit as a class. Additionally, the semi-automatic process for the annotation of a jersey number dataset consisting of 8281 jersey numbers is described. The best recognition rate of 0.83 was achieved by the proposed approach with data augmentation and without using dropout, compared to 0.4 for a more traditional histogram of oriented gradients (HOG) and support vector machine (SVM) based approach.