Using iPhone camera for temporal gait analysis: A case study

Abstract

A comprehensive clinical gait analysis can only be obtained in laboratory setting with motion capture systems. However, it cannot be done at the doctor's office. Hence, affordable, portable and still reliable instruments are being developed, and will allow to obtain some of the gait variables. The Optogait is such an instrument, providing reliable time parameters for gait. Supported by the wide use of smartphones and their potential for sensing applications, we tried to verify if equally reliable data can be obtained from an iPhone6 high speed camera. One healthy participant (22 year-old female) performed several gait cycles, captured by an Optogait system (1000 Hz, Microgate, Italy) and recorded with an iPhone6 camera (240 Hz, Apple Inc, USA) positioned near the ground and 4 m away from the walkway. Nine gait cycles recorded for the left and right sides were analyzed using Kinovea software (www.kinovea.org) and used to calculate stance, swing and double support time (s) from heel strike and toe-off events.

INTRODUCTION

A comprehensive clinical gait analysis can only be obtained in laboratory setting with motion capture systems (Eldeeb & Khodair, 2014; Roriz & Lobo, 2018). However, it cannot be done at the doctor's office. Hence, affordable, portable and still reliable instruments are being developed, and will allow to obtain some of the gait variables. The Optogait is such an instrument, providing reliable time parameters for gait (Lienhard, Schneider, & Maffiuletti, 2013). Supported by the wide use of smartphones and their potential for sensing applications, we tried to verify if equally reliable data can be obtained from an iPhone6 high speed camera.

METHODS

One healthy participant (22 year-old female) performed several gait cycles, captured by an Optogait system (1000 Hz, Microgate, Italy) and recorded with an iPhone6 camera (240 Hz, Apple Inc, USA) positioned near the ground and 4 m away from the walkway. Nine gait cycles recorded for the left and right sides were analysed using Kinovea software (www.kinovea.org) and used to calculate stance, swing and double support time (s) from heel strike and toe-off events. The "iPhone" intra-rater (three raters) and inter-rater agreement along with instruments agreement have been assessed following the procedures recommended by Bland and Altman (1986).

RESULTS

Intra-rater (Fig.1a) and inter-rater agreement (Fig.1b) as well as instruments agreement were excellent (Fig.2).

The mean difference for the "worst" rater was -0.002s±0.006s (PR, Fig.1a). The coefficient of repeatability (0.012s) suggests 95% of the differences will be smaller than 0.012s (<3 frames). The maximum difference found in repeated measurements was 0.020s (5 frames) and it may not have clinical relevance. Results from One-Sample T-test applied to the variable difference [t=0.092, P=0.927 (F); t=-1.941, P=0.058 (PR) and t=0.483, P=0.631 (R)], suggest no evidence for significant (P<0.05) variation in repeated measurements (however, the variable did not follow a normal distribution).

The mean difference between the two raters with lower agreement was 0.000s±0.006s (PR and F; Fig.1b) and between instruments was 0.000s±0.007s (PR and Optogait; Fig. 2).

The One-Sample T-test was applied to the corresponding difference variables (both normally distributed) suggesting no evidence for significant variation between raters [t=0.126, P=0.900>0.05] and instruments [t=0.478, P=0.635>0.05].

CONCLUSIONS

The iPhone6 high-speed camera seems to be a reliable instrument to obtain temporal gait parameters when compared to optogait. An application for quick identification of gait events should be pursued and the study extended for larger samples.

References

References

Bland, J. M., & Altman, D. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. The lancet, 327(8476), 307-310.

Eldeeb, A. M., & Khodair, A. S. (2014). Three-dimensional analysis of gait in postmenopausal women with low bone mineral density. J Neuroeng Rehabil, 11, 55. doi:10.1186/1743-0003-11-55

Lienhard, K., Schneider, D., & Maffiuletti, N. A. (2013). Validity of the Optogait photoelectric system for the assessment of spatiotemporal gait parameters. Medical engineering & physics, 35(4), 500-504.

Roriz, P., & Lobo, A. (2018). Fiber Optical Sensors in Biomechanics. In H. Alemohammad (Ed.), Opto-Mechanical Fiber Optic Sensors (1st ed., pp. 263-300): Elsevier.