Reliability of artificial intelligence-driven markerless motion capture in gait analyses of healthy adults

人工智能驱动的无标记运动捕捉技术在健康成年人步态分析中的可靠性

阅读:1

Abstract

The KinaTrax markerless motion capture system, used extensively in the analysis of baseball pitching and hitting, is currently being adapted for use in clinical biomechanics. In clinical and laboratory environments, repeatability is inherent to the quality of any diagnostic tool. The KinaTrax system was assessed on within- and between-session reliability for gait kinematic and spatiotemporal parameters in healthy adults. Nine subjects contributed five trials per session over three sessions to yield 135 unique trials. Each trial was comprised of a single bilateral gait cycle. Ten spatiotemporal parameters for each session were calculated and compared using the intraclass correlation coefficient (ICC), Standard Error of the Measurement (SEM), and minimal detectable change (MDC). In addition, seven kinematic waveforms were assessed from each session and compared using the coefficient of multiple determination (CMD). ICCs for between-session spatiotemporal parameters were lowest for left step time (0.896) and left cadence (0.894). SEMs were 0.018 (s) and 3.593 (steps/min) while MDCs were 0.050 (s) and 9.958 (steps/min). Between-session average CMDs for joint angles were large (0.969) in the sagittal plane, medium (0.554) in the frontal plane, and medium (0.327) in the transverse plane while average CMDs for segment angles were large (0.860), large (0.651), and medium (0.561), respectively. KinaTrax markerless motion capture system provides reliable spatiotemporal measures within and between sessions accompanied by reliable kinematic measures in the sagittal and frontal plane. Considerable strides are necessary to improve methodological comparisons, however, markerless motion capture poses a reliable application for gait analysis within healthy individuals.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。