Next Article in Journal
Improving GIS-Based Landslide Susceptibility Assessments with Multi-temporal Remote Sensing and Machine Learning
Next Article in Special Issue
Pedestrian Positioning Using a Double-Stacked Particle Filter in Indoor Wireless Networks
Previous Article in Journal
Self-Powered Portable Electronic Reader for Point-of-Care Amperometric Measurements
Previous Article in Special Issue
Logistic Regression for Machine Learning in Process Tomography
Open AccessArticle

Time Coherent Full-Body Poses Estimated Using Only Five Inertial Sensors: Deep versus Shallow Learning

1
Department of Biomedical Signals & Systems, Technical Medical Centre, University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
2
RADiCAL Solutions, LLC. 125 West 31st Street, New York, NY 10001, USA
3
Xsens Technologies B.V., Pantheon 6a, 7521 PR Enschede, The Netherlands
4
Department of Computer Science, Faculty of Electrical Engineering, Mathematics & Computer Science, University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
*
Author to whom correspondence should be addressed.
Matteo Giuberti is with RADiCAL Solutions since March 2019. This work was performed while he was at Xsens Technologies.
Sensors 2019, 19(17), 3716; https://doi.org/10.3390/s19173716
Received: 27 June 2019 / Revised: 30 July 2019 / Accepted: 21 August 2019 / Published: 27 August 2019
(This article belongs to the Special Issue Intelligent Sensor Signal in Machine Learning)
Full-body motion capture typically requires sensors/markers to be placed on each rigid body segment, which results in long setup times and is obtrusive. The number of sensors/markers can be reduced using deep learning or offline methods. However, this requires large training datasets and/or sufficient computational resources. Therefore, we investigate the following research question: “What is the performance of a shallow approach, compared to a deep learning one, for estimating time coherent full-body poses using only five inertial sensors?”. We propose to incorporate past/future inertial sensor information into a stacked input vector, which is fed to a shallow neural network for estimating full-body poses. Shallow and deep learning approaches are compared using the same input vector configurations. Additionally, the inclusion of acceleration input is evaluated. The results show that a shallow learning approach can estimate full-body poses with a similar accuracy (~6 cm) to that of a deep learning approach (~7 cm). However, the jerk errors are smaller using the deep learning approach, which can be the effect of explicit recurrent modelling. Furthermore, it is shown that the delay using a shallow learning approach (72 ms) is smaller than that of a deep learning approach (117 ms). View Full-Text
Keywords: inertial motion capture; machine learning; neural networks; deep learning; LSTM; time coherence; human movement; reduced sensor set; pose estimation inertial motion capture; machine learning; neural networks; deep learning; LSTM; time coherence; human movement; reduced sensor set; pose estimation
Show Figures

Figure 1

MDPI and ACS Style

Wouda, F.J.; Giuberti, M.; Rudigkeit, N.; van Beijnum, B.-J.F.; Poel, M.; Veltink, P.H. Time Coherent Full-Body Poses Estimated Using Only Five Inertial Sensors: Deep versus Shallow Learning. Sensors 2019, 19, 3716.

Show more citation formats Show less citations formats
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop