Annals of biomedical engineering · 2026
Inverse Dynamics Meets Markerless Motion Capture during Standing: Concurrent Validation of Center of Mass and Center of Pressure Estimations.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Nourbakhsh Sabet Dorna, Tamimi Hadi, Vette Albert H, Nazarahari Milad
Аннотация
INTRODUCTION: Observational balance assessment is frequently used clinically but remains subjective and limited in accuracy. Although laboratory-based methods provide precise biomechanical measurements, their need for specialized equipment and expertise limits clinical use. This study presents a multi-view, open-source markerless motion capture system (MMCS) using OpenPose to identify body landmarks, enabling objective balance assessment via center-of-mass (COM) and inverse-dynamics-based center-of-pressure (COP) estimation.
METHODS: The proposed MMCS was experimentally validated. Twelve non-disabled individuals performed five balance tasks (quiet standing on hard and soft surfaces with eyes open and closed, and a weight shift task) against marker-based and force plate measurements. This study also investigated the effect of biomechanical modelling (single-, three-, six-, and eight-segment models) on COM and COP estimations.
RESULTS: Using MMCS, COM trajectories estimated from markerless data showed very high agreement with the marker-based gold standard across models (Pearson's correlation coefficient: CC > 0.90). The COP estimated via the eight-segment model and inverse dynamics showed the highest agreement with force plate data (CC = 0.95-0.98; root-mean-square error: RMSE = 2.2-8.5 mm) in the anterior-posterior direction. Across all models, the anterior-posterior direction showed higher correlations than the medial-lateral direction, except for the weight shift task.
CONCLUSION: Increasing number of segments, improved CC and RMSE across tasks, confirming the value of multi-segment modelling for COM and COP estimation when using MMCS. These findings demonstrate that our MMCS can achieve agreement consistent with laboratory-based systems for balance assessment. By reducing reliance on specialized laboratories while maintaining biomechanical validity, this framework supports balance evaluation in rehabilitation clinics.
METHODS: The proposed MMCS was experimentally validated. Twelve non-disabled individuals performed five balance tasks (quiet standing on hard and soft surfaces with eyes open and closed, and a weight shift task) against marker-based and force plate measurements. This study also investigated the effect of biomechanical modelling (single-, three-, six-, and eight-segment models) on COM and COP estimations.
RESULTS: Using MMCS, COM trajectories estimated from markerless data showed very high agreement with the marker-based gold standard across models (Pearson's correlation coefficient: CC > 0.90). The COP estimated via the eight-segment model and inverse dynamics showed the highest agreement with force plate data (CC = 0.95-0.98; root-mean-square error: RMSE = 2.2-8.5 mm) in the anterior-posterior direction. Across all models, the anterior-posterior direction showed higher correlations than the medial-lateral direction, except for the weight shift task.
CONCLUSION: Increasing number of segments, improved CC and RMSE across tasks, confirming the value of multi-segment modelling for COM and COP estimation when using MMCS. These findings demonstrate that our MMCS can achieve agreement consistent with laboratory-based systems for balance assessment. By reducing reliance on specialized laboratories while maintaining biomechanical validity, this framework supports balance evaluation in rehabilitation clinics.