Gait Mimicking - Attack Resistance Testing of Gait Authentication Systems

Recent advances in technology have shown us that we can automatically recognize individuals based on the way they walk. However, gait biometrics is a rather new area within biometrics, so the security of such technology needs to be challenged. We present a study that focus on imitation, or mimicking of gait. The bottom line question is whether it is possible to learn to walk like someone else. If this would turn out to be easy, it will have a severe effect on the potential of gait as an authentication mechanism in the future. We have developed a software tool that uses wearable sensors to collect and analyze gait acceleration data. The research is further based on an experiment, involving extensive training of test subjects, and using various sources of feedback like video and statistical analysis. The attack scores are analyzed by regression, and the goal is to determine whether or not the participants are increasing their mimicking skills, or simply put: if they are learning. The first part of the experiment is a ''friendly'' scenario, involving 50 participants that are successfully enrolled into a gait authentication system. The results compete with state of the art gait technology, with an EER of 6.2%. The rest of the experiment is related to mimicking, and the research discovers that six out of seven participants seem to have a natural boundary to their performance. A very limited amount of learning is present, not nearly enough to pose a threat to gait biometrics. The findings suggest that gait mimicking is a very difficult task, and our physiological characteristics work against us when we try to imitate someone elses walking. In most cases, training for it will \textit{worsen} the impostor's performance