A video-based assessment tool using machine learning for ergonomic risk prediction in manual lifting tasks
Kung-Jeng Wang1,2, Kim-Tien Truong1
1Department of Industrial Management, National Taiwan University of Science and Technology, Taipei, Taiwan, ROC.
Ergonomics
|December 27, 2025
Summary
This study introduces a video-based tool for assessing ergonomic risks in manual lifting tasks. It quantifies risks using the NIOSH lifting equation, offering a cost-effective solution for industries.
Area of Science:
- Occupational health and safety
- Ergonomics
- Biomechanics
Background:
- Manual lifting is prevalent in logistics, healthcare, and manufacturing.
- These tasks pose significant ergonomic risks, leading to musculoskeletal disorders.
Purpose of the Study:
- To develop a video-based ergonomic risk assessment tool.
- To quantify risks using the revised NIOSH lifting equation from smartphone videos.
Main Methods:
- Integrated MediaPipe Pose Landmarker for joint tracking.
- Employed classification and regression trees for lifting stage classification.
- Utilized a back-propagation neural network for parameter compensation.
Main Results:
- The tool calculates Recommended Weight Limit (RWL) and Lifting Index (LI).
- Validated on laboratory and field datasets.
- Demonstrated adaptability, scalability, and cost-effectiveness.
Conclusions:
- The proposed tool offers a practical method for ergonomic risk assessment in manual lifting.
- It provides a cost-effective and accessible solution for industries.
- Potential to reduce musculoskeletal disorders through improved ergonomic practices.
