EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

CLUMM: Contrastive Learning for Unobtrusive Motion Monitoring

Pius Gyamenah, Hari Iyer iD, Heejin Jeong iD, Shenghan Guo iD

Sensors · 2025

DOI: 10.3390/s25041048 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, 2025. Verfasst von Pius Gyamenah und 3 weiteren Autor:innen. Beteiligte Einrichtung: The School of Manufacturing Systems and Networks, Ira A. Fulton Schools of Engineering, Arizona State University, Mesa, AZ 85212, USA. Hauptthemen laut MEDLINE: Machine Learning. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 70 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NIOSH CDC HHS, National Institute for Occupational Safety and Health (NIOSH), Arizona State University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Traditional approaches for human monitoring and motion recognition often rely on wearable sensors, which, while effective, are obtrusive and cause significant discomfort to workers. More recent approaches have employed unobtrusive, real-time sensing using cameras mounted in the manufacturing environment. While these methods generate large volumes of rich data, they require extensive labeling and analysis for machine learning applications. Additionally, these cameras frequently capture irrelevant environmental information, which can hinder the performance of deep learning algorithms. To address these limitations, this paper introduces a novel framework that leverages a contrastive learning approach to learn rich representations from raw images without the need for manual labeling. This framework mitigates the effect of environmental complexity by focusing on critical joint coordinates relevant to manufacturing tasks. This approach ensures that the model learns directly from human-specific data, effectively reducing the impact of the surrounding environment. A custom dataset of human subjects simulating various tasks in a workplace setting is used for training and evaluation. By fine-tuning the learned model for a downstream motion classification task, we achieve up to 90% accuracy, demonstrating the effectiveness of our proposed solution in real-time human motion monitoring.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningHumansAlgorithmsMotionWearable Electronic DevicesMovementMonitoring, PhysiologicDeep Learning

Stichwörter der Autor:innen

contrastive learningin situ monitoringmotion recognitionself-supervised learningunobtrusive human sensing

Autor:innen und Institutionen

  • The School of Manufacturing Systems and Networks, Ira A. Fulton Schools of Engineering, Arizona State University, Mesa, AZ 85212, USA.
  • The Polytechnic School, Ira A. Fulton Schools of Engineering, Arizona State University, Mesa, AZ 85212, USA.

ORCID-Profile: Hari Iyer, Heejin Jeong, Shenghan Guo

Förderung

  • NIOSH CDC HHS
  • National Institute for Occupational Safety and Health (NIOSH)
  • Arizona State University

Zitieren

APA

Gyamenah, P., Iyer, H., Jeong, H., & Guo, S. (2025). CLUMM: Contrastive Learning for Unobtrusive Motion Monitoring. Sensors. https://doi.org/10.3390/s25041048

Vancouver

Gyamenah P, Iyer H, Jeong H, Guo S. CLUMM: Contrastive Learning for Unobtrusive Motion Monitoring. Sensors. 2025. doi:10.3390/s25041048

BibTeX
@article{Gyamenah2025,
  author = {Gyamenah, Pius and Iyer, Hari and Jeong, Heejin and Guo, Shenghan},
  title = {CLUMM: Contrastive Learning for Unobtrusive Motion Monitoring},
  journal = {Sensors},
  year = {2025},
  doi = {10.3390/s25041048}
}
Kostenlos starten →