TY - JOUR
T1 - Internet-of-Things-Enabled Markerless Running Gait Assessment from a Single Smartphone Camera
AU - Young, Fraser
AU - Mason, Rachel
AU - Morris, Rosie
AU - Stuart, Samuel
AU - Godfrey, Alan
N1 - Funding information: This work was supported by the European Regional Development Intensive Industrial Innovation Programme (IIIP) as part of doctoral research, Grant Number: 25R17P01847. Rachel Mason is co-funded by DANU sports and the faculty of health and life sciences, Northumbria University. Dr Stuart is supported, in part, by funding from the Parkinson’s Foundation (PF-FBS-1898, PFCRA-2073).
PY - 2023/1/7
Y1 - 2023/1/7
N2 - Running gait assessment is essential for development oftechnique optimization strategies as well as to inform injury prevention andrehabilitation. Currently, running gait assessment relies on (i) visualassessment, exhibiting subjectivity and limited reliability, or (ii) use ofinstrumented approaches, which often carry high costs and can be intrusive dueto attachment of equipment to the body. Here use of an IoT-enabled markerlesscomputer vision smartphone application based upon Googles pose estimation modelBlazePose was evaluated for running gait assessment for use in low-resourcesettings. That human pose estimation architecture was used to extract contacttime, swing time, step time, knee flexion angle and foot strike location from alarge cohort of runners. The gold-standard Vicon 3D motion capture system wasused as a reference. The proposed approach performs robustly, demonstratinggood (ICC(2,1) > 0.75) to excellent (ICC(2,1) >0.90) agreement in all running gait outcomes. Additionally, temporal outcomesexhibit low mean error (0.01-0.014s) in left foot outcomes. However, there aresome discrepancies in right foot outcomes, due to occlusion. This studydemonstrates that the proposed low-cost and markerless system provides accuraterunning gait assessment outcomes. The approach may help routine running gaitassessment in low-resource environments.
AB - Running gait assessment is essential for development oftechnique optimization strategies as well as to inform injury prevention andrehabilitation. Currently, running gait assessment relies on (i) visualassessment, exhibiting subjectivity and limited reliability, or (ii) use ofinstrumented approaches, which often carry high costs and can be intrusive dueto attachment of equipment to the body. Here use of an IoT-enabled markerlesscomputer vision smartphone application based upon Googles pose estimation modelBlazePose was evaluated for running gait assessment for use in low-resourcesettings. That human pose estimation architecture was used to extract contacttime, swing time, step time, knee flexion angle and foot strike location from alarge cohort of runners. The gold-standard Vicon 3D motion capture system wasused as a reference. The proposed approach performs robustly, demonstratinggood (ICC(2,1) > 0.75) to excellent (ICC(2,1) >0.90) agreement in all running gait outcomes. Additionally, temporal outcomesexhibit low mean error (0.01-0.014s) in left foot outcomes. However, there aresome discrepancies in right foot outcomes, due to occlusion. This studydemonstrates that the proposed low-cost and markerless system provides accuraterunning gait assessment outcomes. The approach may help routine running gaitassessment in low-resource environments.
U2 - 10.3390/s23020696
DO - 10.3390/s23020696
M3 - Article
VL - 23
JO - Sensors
JF - Sensors
SN - 1424-3210
IS - 2
M1 - 696
ER -