Using Artificial Intelligence for Pattern Recognition in a Sports Context.

Ana Cristina Nunes Rodrigues,Rui Manuel Sousa Mendes,António José Figueiredo,Alexandre Santos Pereira,André Gonçalves Araújo,Micael Santos Couceiro

doi:10.3390/s20113040

Abstract

Optimizing athlete’s performance is one of the most important and challenging aspects of coaching. Physiological and positional data, often acquired using wearable devices, have been useful to identify patterns, thus leading to a better understanding of the game and, consequently, providing the opportunity to improve the athletic performance. Even though there is a panoply of research in pattern recognition, there is a gap when it comes to non-controlled environments, as during sports training and competition. This research paper combines the use of physiological and positional data as sequential features of different artificial intelligence approaches for action recognition in a real match context, adopting futsal as its case study. The traditional artificial neural networks (ANN) is compared with a deep learning method, Long Short-Term Memory Network, and also with the Dynamic Bayesian Mixture Model, which is an ensemble classification method. The methods were used to process all data sequences, which allowed to determine, based on the balance between precision and recall, that Dynamic Bayesian Mixture Model presents a superior performance, with an F1 score of 80.54% against the 33.31% achieved by the Long Short-Term Memory Network and 14.74% achieved by ANN.

Highlights

Pattern recognition is the ability of a machine to recognize patterns in their environment, using artificial intelligence learning abilities to make decisions
The current Section shows the results of both action recognition approaches
This led us to conclude that the fifth test is the one with better results achieved, implying that all features should be considered for further analysis and to compare artificial neural networks (ANN), Long Short-term Memory (LSTM) and

Summary

Introduction

Pattern recognition is the ability of a machine to recognize patterns in their environment, using artificial intelligence learning abilities to make decisions. Several works have compared the performance of wearable-based systems with professional laboratory/clinic equipment and verified that, while results were rather similar, wearable solutions were able to keep the ecological validity of task [2] This implies that the data extracted with wearable solutions shall represent ecological patterns of the human movement inherent to various areas of application, such as sports, speech and face recognition, classification of handwritten characters, medical diagnosis and rehabilitation [3]. Wearable technologies allow to extract, on-the-fly and in a real-world context, kinematic variables, such as the position, velocity, orientation, and others, and physiological data, such as electromyography (EMG) and heart rate It is still unclear how this large amount of data can be translated into simple and practical data. This is considerably more challenging in the context of sports, where coaches, sports analysts, exercise physiologists and athletes need to have access to this data in real-time, to improve the understanding of the athlete’s behaviour during training and competition [4]

Objectives

Methods

Results

Discussion

Conclusion