New area based metrics for gait recognition

Abstract

Gait is a new biometric aimed to recognise a subject by the manner in which they walk. Gait has several advantages over other biometrics, most notably that it is a non-invasive and perceivable at a distance when other biometrics are obscured. We present a new area based metric, called gait masks, which provides statistical data intimately related to the gait of the subject. Early results show promising results with a recognition rate of 90% on a small database of human subjects. In addition to this, we show how gait masks can also be used on subjects other than humans to provide information about the gait cycle of the subject. © Springer-Verlag 2001.