Appropriateness of gait analysis for biometrics: Initial study using FDA method

Abstract

Human body movement has been under continuous research for many years due to its potential application as a novel biometric system to identify individuals. It is possible to utilize various techniques, not only to obtain requested movement data, but also to analyse movement data. This paper uses functional data analysis on data obtained from 12 volunteers and uses 20 markers from the 3D motion capture system VICON MX T020. The functional data analysis was chosen as a suitable tool to obtain more information about an individual's movement because it uses a technique for real-time data, which corresponds to continuous time process. The results show that all markers, under any walking speed and condition, identify a significantly high percentage of individual pairs. Further, our results discriminate between markers, where some markers are highly dependent on walking speed and condition, and also on the influence of body part asymmetry. In addition, regular movement patterns in almost all participants’ data shows a potential to identify individuals based on gait recognition with a 1:1 matching result. © 2017 Elsevier Ltd