Optimising computer vision-based ergonomic assessments: sensitivity to camera position and monocular 3D pose model

Numerous computer vision algorithms have been developed to automate posture analysis and enhance the efficiency and accuracy of ergonomic evaluations. However, the most effective algorithm for conducting ergonomic assessments remains uncertain. Therefore, the aim of this study was to identify the optimal camera position and monocular 3D pose model that would facilitate precise and efficient ergonomic evaluations. We evaluated and compared four currently available computer vision algorithms: Mediapipe BlazePose, VideoPose3D, 3D-pose-baseline, and PSTMO to determine the most suitable model for conducting ergonomic assessments. Based on the findings, the side camera position yielded the lowest Mean Absolute Error (MAE) across static, dynamic, and combined tasks. This positioning proved to be the most reliable for ergonomic assessments. Additionally, VP3D_FB demonstrated superior performance among evaluated models. Practitioner Summary: This study aimed to determine the most effective computer vision algorithm and camera position for precise and efficient ergonomic evaluations. Evaluating four algorithms, we found that the side camera position with VideoPose3D yielded the lowest Mean Absolute Error (MAE), ensuring precise and efficient evaluations.