System for classification of human gaits using markerless motion capture sensor

Abstract

Purpose

Human gait analysis is based on a significant part of the musculoskeletal, nervous and respiratory systems. Gait analysis is widely adopted to help patients increase community involvement and independent living.

Design/methodology/approach

This paper presents a system for the classification of abnormal human gaits using a Markerless 3D Motion Capture device. This study aims at examining and estimating the spatiotemporal and kinematic parameters obtained by 3D gait analysis in diverse groups of gait-impaired subjects and compares the parameters with that of healthy participants to interpret the gait patterns.

Findings

The classification is based on mathematical models that distinguish between normal and abnormal gait patterns depending on the deviations in the gait parameters. The difference between the gait measures of the control and each disease group was examined using 95% limits of agreement by the Bland and Altman method. The scatter plots demonstrated gait variability in Parkinsonian and ataxia gait and knee joint angle variation in hemiplegic gait when compared with those of healthy controls. To prove the validity of the Kinect camera, significant correlations were detected between Kinect- and inertial-based gait tests.

Originality/value

The various techniques used for gait assessments are often high in price and have existing limitations like the hindrance of components. The results suggest that the Kinect-based gait assessment techniques can be used as a low-cost, less-intrusive alternative to expensive infrastructure gait lab tests in the clinical environment.

Citation

Madhana, K., Jayashree, L.S. and Perumal, K. (2023), "System for classification of human gaits using markerless motion capture sensor", Journal of Enabling Technologies, Vol. 17 No. 2, pp. 41-53. https://doi.org/10.1108/JET-08-2022-0058