Collaboration in computation at Curtin University: Investigating human activity recognition and joint kinematics in ballet

Dr Kathryn Napier1, Dr Kevin Chai1, Ms Danica Hendry2, Dr Richard Hosking1, Dr Amity Campbell2, Prof Leon Straker2, Dr Luke Hopper3, Prof Tele Tan4, Professor Peter O’Sullivan2

1Curtin Institute For Computation, Curtin University, Perth, Australia, 2School of Physiotherapy and Exercise Science, Curtin University, Perth, Australia, 3Western Australian Academy of Performing Arts, Edith Cowan University, Perth, Australia, 4School of Civil and Mechanical Engineering, Curtin University, Perth, Australia

Abstract:

Introduction

The Curtin Institute for Computation (CIC) and the School of Physiotherapy and Exercise Sciences have collaborated to develop a novel machine learning based approach to accurately measure ballet dancer training load and joint kinematics in a real-world training setting. Currently training load is estimated from dancer’s written diary entries, while the investigation of joint kinematics requires sophisticated and expensive laboratory based optical motion capture systems. By combining expertise in both physiotherapy and computation, we have developed a novel methodology utilising affordable wearable sensors that can accurately measure dancer training load and joint kinematics in real-world ballet classes.

Methods

Two separate studies were performed on female ballet dancers fitted with wearable sensors. The first study developed a human activity recognition (HAR) convolutional neural network (CNN) model to classify six different jumping and leg lift ballet movements. The second study developed a joint kinematics recurrent neural network (RNN) prediction model for leg lift movements predicting thigh angle as a measurement of leg height.

Results

The HAR CNN model achieved 83% classification accuracy, and the joint kinematics RNN model achieved peak thigh angle predictions with a mean error of 5.9% based on the results obtained from the experimental datasets.

Conclusion

The models developed were robust enough to identify jumping and leg lifting movements and to identify how often and how high dancers lifted their legs during leg lifts in a real-world ballet training class. This methodology will assist in providing further insight into the factors influencing a dancer’s pain and injury risk.


Biography:

Kathryn Napier is a Senior Data Scientist with the Curtin Institute for Computation (CIC) at Curtin University. The CIC was founded in 2015 to initiate and foster collaborative, interdisciplinary research that applies computational methods. Kathryn collaborates with Curtin University researchers and project partners to assist with data, computational, analytics and visualisation problems. Prior to joining the Curtin Institute for Computation in late 2018, Kathryn worked as a Research Associate at the Centre for Comparative Genomics at Murdoch University in the fields of Bioinformatics and Health Informatics

ABOUT AeRO

AeRO is the industry association focused on eResearch in Australasia. We play a critical coordination role for our members, who are actively transforming research via Information Technology. Organisations join AeRO to advance their own capabilities and services, to collaborate and to network with peers. AeRO believes researchers and the sector significantly benefit from greater communication, coordination and sharing among the increasingly different and evolving service providers.

Conference Managers

Please contact the team at Conference Design with any questions regarding the conference.
© 2019 Conference Design Pty Ltd