Recognising Care-Related Activities with a Convolutional Neural Network Using Inertial Measurement Units

Fenja T. Hesselmann, Jannik Fleßner, Alexander Pauls, Vanessa Cobus, Frauke Koppelin, Frank Wallhoff

2025

Abstract

Sensor-based human activity recognition is a growing field of research. In addition to recognising everyday movements, situation-dependent activities can also be detected. This paper therefore aims to detect care-specific movements. For this purpose, 13 different nursing activities were recorded with Inertial Measurement Units (IMUs) worn on the body. In this paper, we present an approach on how the sensor data can be used for recognition. Convolutional neural networks were used for classification. The focus of this work is on two different fusion approaches of the data to check which approach achieves better results. In the first approach, all data is fused at the beginning, while in the second one, a separate pipeline is designed for each sensor and fused later. The results show that a later fusion technique provides a better F1 score of 90.2 % compared to a model that considers all signals from the beginning (F1 score: 82.5 %).

Paper Citation

in Harvard Style

Hesselmann F., Fleßner J., Pauls A., Cobus V., Koppelin F. and Wallhoff F. (2025). Recognising Care-Related Activities with a Convolutional Neural Network Using Inertial Measurement Units. In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF; ISBN 978-989-758-731-3, SciTePress, pages 727-733. DOI: 10.5220/0013257400003911

in Bibtex Style

@conference{healthinf25,
author={Fenja Hesselmann and Jannik Fleßner and Alexander Pauls and Vanessa Cobus and Frauke Koppelin and Frank Wallhoff},
title={Recognising Care-Related Activities with a Convolutional Neural Network Using Inertial Measurement Units},
booktitle={Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF},
year={2025},
pages={727-733},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013257400003911},
isbn={978-989-758-731-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies - Volume 2: HEALTHINF
TI - Recognising Care-Related Activities with a Convolutional Neural Network Using Inertial Measurement Units
SN - 978-989-758-731-3
AU - Hesselmann F.
AU - Fleßner J.
AU - Pauls A.
AU - Cobus V.
AU - Koppelin F.
AU - Wallhoff F.
PY - 2025
SP - 727
EP - 733
DO - 10.5220/0013257400003911
PB - SciTePress