Assessment of spatiotemporal gait parameters using a deep learning algorithm-based markerless motion capture system

Abstract

Spatiotemporal parameters can characterize the gait patterns of individuals, allowing assessment of their health status and detection of clinically meaningful changes in their gait. Video-based markerless motion capture is a user-friendly, inexpensive, and widely applicable technology that could reduce the barriers to measuring spatiotemporal gait parameters in clinical and more diverse settings. Two studies were performed to determine whether gait parameters measured using markerless motion capture demonstrate concurrent validity with those measured using marker-based motion capture and pressure sensitive gait mats. For the first study, thirty healthy adults performed treadmill gait at self-selected speeds while marker-based motion capture and synchronized video data were recorded simultaneously. For the second study, twenty-five healthy adults performed over-ground gait at self-selected speeds while footfalls were recorded using a gait mat and synchronized video data were recorded simultaneously. Kinematic heelstrike and toe-off gait events were used to identify the same gait cycles between systems. Nine spatiotemporal gait parameters were measured by each system and directly compared between systems.Measurements were compared using Bland-Altman methods, mean differences, Pearson correlation coefficients, and intraclass correlation coefficients. The results indicate that markerless measurements of spatiotemporal gait parameters have good to excellent agreement with marker-based motion capture and gait mat systems, except for stance time and double limb support time relative to both systems and stride width relative to the gait mat. These findings indicate that markerless motion capture can adequately measure spatiotemporal gait parameters during treadmill and overground gait.

Applications and limitations of current markerless motion capture methods for clinical gait biomechanics

“…Standard deviations of lower limb joint angles were between 3 and 10 degrees with the markerless method compared to the marker-based method, and correlations for hip and ankle frontal and rotation planes were poor (0.26–0.51), indicating high variability of this system. Most recently, Theia3D markerless software (Theia Markerless Inc.) which uses a proprietary pose estimation algorithm was compared between an 8 camera markerless system (85 Hz) and a seven camera marker-based system (85 Hz) ( Kanko et al, 2021b ; Kanko et al, 2021c ). They reported no bias or statistical difference for walking spatial measures ( e.g.…”

“…, step length, step width, velocity) and a small difference in temporal measures ( e.g. , swing time and double support time) ( Kanko et al, 2021c ). A follow-on study using the same data found average differences of 22–36 mm for joint centers and 2.6–11 degrees for flexion/extension and abduction/adduction, although rotation about the longitudinal axis differences were 6.9–13.2 degrees compared to marker-based methods ( Kanko et al, 2021b ).…”

“…Importantly, the lower ranges of these translational and rotational differences are within error rates identified by previous research ( Fiorentino et al, 2017 ; Kessler et al, 2019 ; Miranda et al, 2013 ). These strong results appear to be due to Theia3D having labelled their own biomechanically applicable data set which identifies 51 keypoints on the body ( Kanko et al, 2021b ; Kanko et al, 2021c ), compared to OpenPose which only identifies 25 points ( Cao et al, 2018 ). However, Theia3D software is somewhat of a black box, as it is unknown exactly which keypoints are being used.…”

Concurrent assessment of gait kinematics using marker-based and markerless motion capture

“…The data collection setup and procedure were previously used to compare spatiotemporal gait parameter measurements between marker-based and markerless motion capture (Kanko et al, 2021b).…”

Analysis of Volleyball Video Intelligent Description Technology Based on Computer Memory Network and Attention Mechanism

“…If the action in volleyball video needs to be extracted by intelligent analysis and description, it needs to be realized by intelligent algorithm in image recognition technology [ 26 ]. With the development of artificial intelligence technology, in terms of data processing of video information, image recognition technology based on deep learning algorithm has become a research hotspot and development trend in the fields of video and image information mining and data association analysis [ 27 , 28 ]. What video intelligent description technology needs to solve is to divide the target data set in the image or video group into multiple groups according to a certain correlation law.…”