Abstract
Background: Openpose is a motion capture system that uses artificial intelligence. This study aimed to verify the reliability and validity of motion angular velocity using Openpose in trunk flexion-extension movements (TFEM) and compare Openpose with methods that use accelerometers and marker measurements. Methods: Fifty healthy young participants were included in the study. The TFEM tasks were performed. Motion analyses were performed using accelerometers, markers, and the Openpose software. To examine the reliability and validity of the angular velocity of motion, we calculated intraclass correlation coefficients (ICC) (1,1) and ICC (1,3) for each maximum flexion and extension angular velocity. In addition, we performed cross-lag correlation analyses to examine similarities between the analysis methods. Results: All the ICC analysis methods were “almost perfect.” In addition, the cross-lag correlation analysis showed a high correlation among all the analysis methods. Conclusion: In this study, the motion analysis method using Openpose demonstrated high reliability and validity of measurements of angular velocity of motion in TFEM. Therefore, the Openpose measurement is helpful for objectively assessing motor control in TFEM.
Recommended Citation
Katayama, Joga; Tennan, Kai; Harada, Sakino; Hirata, Aoi; Shigetoh, Hayato; and Miyazaki, Junya
(2025)
"Reliability and validity of trunk flexion-extension angular velocity measurement in healthy young adults using AI-enhanced motion capture,"
Asian Journal of Physical Therapy: Vol. 2:
Iss.
2024, Article 3.
Available at:
https://digitalcommons.lmunet.edu/ajpt/vol2/iss2024/3
Submit Article