Ergonomic Assessment Using Human Pose Estimation: A Real-Time Approach with YOLO and BlazePose

• John Davi D. C. Pires UFAL
• Marcelo Costa Oliveira UFAL
• Baldoíno Fonseca UFAL
• Márcio Ribeiro UFAL
• Bruno Antônio F. Giglio SESI-AL
• Alexandre Xavier Calzado SESI-AL

DOI: https://doi.org/10.5753/sbcas_estendido.2025.7623

Resumo

Musculoskeletal disorders (MSDs) are a major cause of workplace injuries, often linked to poor posture and repetitive movements. Traditional assessments like RULA are subjective and labor-intensive. To overcome these limitations, this study employs computer vision for automated ergonomic evaluation. Using YOLO for person detection and BlazePose for pose estimation, the system analyzes workplace postures via surveillance cameras, generating RULA scores to assess ergonomic risks. Tested on the MPI-INF-3DHP dataset, it achieves MPJPE-PA values between 243.91 mm and 295.70 mm. This scalable, objective approach enhances workplace safety. Future work will refine accuracy, handle real-world variations, and integrate real-time feedback.

Referências

Abobakr, A., Nahavandi, D., Hossny, M., Iskander, J., Attia, M., Nahavandi, S., and Smets, M. (2019). Rgb-d ergonomic assessment system of adopted working postures. Applied Ergonomics, 80:75–88. Epub 2019 May 25.

Chidambaram, V., Mohan, G., .M, V., and Kanchan, B. (2023). Ergonomic investigations on novel dynamic postural estimator using blaze pose and transfer learning. Ergonomics, 67:1–35.

ErgoPlus (2024). Definition of musculoskeletal disorder (msd). [link].

Gong, W., Zhang, X., Gonzàlez, J., Sobral, A., Bouwmans, T., Tu, C., and Zahzah, E.-h. (2016). Human pose estimation from monocular images: A comprehensive survey. Sensors, 16(12).

Ionescu, C., Papava, D., Olaru, V., and Sminchisescu, C. (2014). Human3.6m: Large scale datasets and predictive methods for 3d human sensing in natural environments. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(7):1325–1339.

Khanam, R. and Hussain, M. (2024). Yolov11: An overview of the key architectural enhancements.

Li, L. and Xu, X. (2019). A deep learning-based rula method for working posture assessment. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 63(1):1090–1094.

Mehta, D., Rhodin, H., Casas, D., Fua, P., Sotnychenko, O., Xu, W., and Theobalt, C. (2017). Monocular 3d human pose estimation in the wild using improved cnn supervision. In 3D Vision (3DV), 2017 Fifth International Conference on. IEEE.

Nayak, G. K. and Kim, E. (2021). Development of a fully automated rula assessment system based on computer vision. International Journal of Industrial Ergonomics, 86:103218.

Oliveira, P. G. d. and Santana, V. S. (2024). Space-time analysis of work-related musculoskeletal disorders in brazil: an ecological study. Cadernos de Saúde Pública, 40(7):e00141823.

Plantard, P., Shum, H. P., Le Pierres, A.-S., and Multon, F. (2017). Validation of an ergonomic assessment method using kinect data in real workplace conditions. Applied Ergonomics, 65:562–569.

Ramírez, S. (2024). Fastapi. [link].

Roman-Liu, D. (2014). Comparison of concepts in easy-to-use methods for msd risk assessment. Applied Ergonomics, 45(3):420–427.

Zheng, C., Wu, W., Chen, C., Yang, T., Zhu, S., Shen, J., Kehtarnavaz, N., and Shah, M. (2023). Deep learning-based human pose estimation: A survey.