Southampton University investigates the biometric potential of OAE

The security industry finds itself in a no-win situation. Consumers demand quick and instant access to resources, but appear reluctant to cooperate with invasive or time-consuming identity verification. But scientists from Southampton University are aiming to change that by investigating the potential for Otoacoustic emissions (OAE) to be used as a biometric application for identification purposes. The unique three year project, which began in January 2007, is conducted by researchers from the School of Electronics and Computer Sciences is set to create a new-look personal identity verification schema. All for a research grant of just three-hundred and fifty thousand pounds – a bargain when you consider identity derived fraud reportedly costs the government nearly two billion pounds every year. So how does it work?

Well, OAE are low intensity sounds produced by our ears in response to audio stimulation and are generated by the activity of the outer hair cells. The emissions produced by the human process of amplifying low level sounds can be detected at the entrance of the ear. Evidence suggests that OAE is unique to each individual and can even be used to distinguish gender and ethnicity. Not only have OAE proven to be exclusive to a person, the characteristics depend upon the input sound in a manner that also varies between individuals. This offers particular opportunities when applying the practice as a biometric system.

Leading the investigation is Dr. Steve Beeby. He said, “We hope the project will establish OAE as a robust biometric analysis which in the future will be used to identify and verify individuals. If successful, the schema would have a significant impact on the security industry. Because a high level of classification performance can be obtained using the raw time-pressure data, the potential is there to satisfy consumer and producer demands.”

There are a number of advantages an otoacoustic-based biometric system enjoys over other forms of identity verification. Firstly, it can be embodied in a telephone handset or set of headphones, technology which everybody is familiar with and thus is socially acceptable.

Secondly, OAE can be employed in a challenge-response dialogue (whereby one party presents a question and another party must provide a valid answer such as a password to be authenticated.) The stimulus dependent nature of the OAE will increase the performance and effectiveness of the biometric. In addition, OAEs can potentially help retrieve stolen mobile phones by rendering them useless if the phone recognises the user is not the legitimate owner and subsequently disables itself.

However, there are still elements that need investigating if OAE is ever going to usurp other identity verification options. Dr Beeby said, “The potential long term drift of emissions, the influence of hearing impairments such as ear infections and the effect of external noise are issues which could affect the reliability of our results. Also, in subjects who have consumed alcohol or drugs, the emissions can be deadened or altered.

There’s still a while before OAE will become a fixture in real life security applications. First the technique will have to consistently churn out low false-match rates and prove that an individual’s recorded OAE stays recognisable over a long term period. Dr Beeby said, “There’s a lot still to do but we’re working in conjunction with the Engineering and Physical Sciences Research Council (EPSRC). The first step is developing an on-the-ear probe that efficiently captures the emissions in a range of situations, and is acceptable to the user. We’re confident that in five to ten years the OAE biometric system will be a sufficiently developed application.”

(Smartcard News Ltd, 2009)