Aryo Nugroho, Nurwahyu Alamsyah, Rosita Meitha Surjani, Ni Kadek Dessy Hariyanti, Ardi Ilham Falah, I Nyoman Gede Arya Astawa
The growing shift towards hybrid and remote work has intensified the risk of work-related musculoskeletal disorders (WMSDs) caused by prolonged sedentary behavior. Traditional ergonomic assessments, while effective, are limited by their static nature and lack of scalability for real-time monitoring. This study presents a lightweight, camera-based machine learning framework for classifying seated postures into ergonomic and non-ergonomic categories in real-time. Utilizing consumer-grade smartphone cameras and MediaPipe Pose, we extract six structured ergonomic features - neck angle, back angle, arm angle, leg angle, shoulder asymmetry, and sitting duration. These features are evaluated against ISO 11226 and NIOSH thresholds and labeled by certified ergonomics experts. We compare three supervised classifiers - Support Vector Machine, Random Forest, and Multi-Layer Perceptron - for their classification performance and interpretability. The Random Forest model achieved the highest accuracy (92 %) and provided valuable insights into feature importance. This interpretable and scalable solution contributes to the vision of sustainable digital ergonomics by enabling non-intrusive posture monitoring and supporting health-aware smart work environments. © 2025 IEEE.
Universitas Muhammadiyah Yogyakarta, Universitas PGRI Adi Buana Surabaya Research Fellow, Department of Electrical Engineering, Yogyakarta, Indonesia; Universitas Muhammadiyah Yogyakarta, Department of Information Technology, Yogyakarta, Indonesia; University of Surabaya, Department of Industrial Engineering, Surabaya, Indonesia; Politeknik Negeri Bali, Department of Business Administration, Bali, Denpasar, Indonesia; Institut Teknologi Sepuluh Nopember, Department of Urban and Regional Planning, Surabaya, Indonesia; Politeknik Negeri Bali, Deparment of Information Technology, Denpasar, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock