Otoacoustic Emission Based Biometric Systems

点击翻译按钮获取中文摘要

Identity-derived fraud is reported to cost the UK 1.72 billion per year. The security industry is caught between two mutually opposing pressures: consumers demand ever more and instant access to their resources, whilst at the same time remaining unwilling to cooperate with invasive and/or time-consuming identity verification. The challenge, then, is to create a personal identity verification schema that is highly reliable, mom-invasive, fast, socially acceptable and cheap to deploy. Biometrics, in the context of this proposal, refers to the quantitative measurement of the physiological or behavioural attributes of an individual with a view to confirming (or not) or establishing their identity. Otoacoustic emissions (OAE) are noises emitted by the mammalian ear in response to an audio stimulation, and a growing body of evidence suggests that these responses can be used to distinguish individuals. The source of OAE lies with the human body's process for amplifying low level sounds. The emissions produced by this process can be detected at the entrance to the ear and are typically evoked by stimulating the hearing process by a suitable input sound. Not only have OAE been shown to be unique to individuals, their characteristics depend upon the input sound in a manner that also varies between individuals. This fact offers some unique opportunities when applying this as a biometric system.An otoacoustic-based biometric system enjoys a number of significant advantages over other identity verification schemas:- It can be embodied is a telephone handset or headphones - everybody is familiar with these.- It can be employed in a challenge-response dialogue: The stimulus dependant nature of the OAE offers the potential for a challenge-response approach hidden from the subject greatly increasing the performance of the biometric, especially with respect to circumvention.In summary, this proposal is aimed at establishing OAE as a robust biometric characteristic. An on-the-ear probe acceptable to the users will be developed that efficiently excites and captures the emissions. The biometric analysis used to identify/verify individuals will be optimised and the system performance measured using standard metrics. Practicalities such as potential long term drift of the emissions, the influence of hearing impairments and external noise will be investigated to evaluate the potential for using OAE in real applications. Finally, the challenge response dialogue will be applied and its effect on the biometric performance of the system evaluated.