The Shakespeare experiment: Preliminary results for the recognition of a person based on the sound of walking

Abstract

Gait Recognition is a behavioural biometric that has been investigated for a few decades now. Various sensors can be used in this research. Most research is using video capture and various methods for analyzing the video. This method is well suited in case of user identification and has proven its value in court cases as forensic evidence. Another way to do gait recognition is using wearable sensors, most often accelerometers, which are available in any modern smart phone. In this case, the collected acceleration data can be matched against a profile of the gait of the owner of the phone and it can be established if the person carrying the phone is the owner or not. Less gait recognitions research has been done using other kinds of sensors. For example, floor sensors have been used, creating a situation where for example an entrance to a building or part of a building can be secured by measure the ground reaction force of someone walking on this specially equipped floor. As was noted by Shakespeare already (e.g. in The Tempest he writes: Highst Queen of state, Great Juno comes; I know her by her gait), can people also be recognized by the sound of their walking. In this paper, we will describe our preliminary results of analyzing the recorded sound of a person walking. As an homage to Shakespeare for “initiating” the field of biometric sound-based gait recognition over 400 years ago, has our experiment been termed “The Shakespeare Experiment (TSE)”. The participants in this experiment wore two different types of shoes, where the soles were significantly different to provide distinct differences in the sound. We have analyzed our data both in a same-shoe and a cross-shoe setting. In an identification setting we obtained an initial recognition accuracy of nearly 80%.