Good Thinking

Earbud tech identifies users by the inside of their ears

Earbud tech identifies users b...
Plans call for the EarEcho system to work in either of the user's earbuds, or perhaps in both at once for better accuracy
Plans call for the EarEcho system to work in either earbud, or perhaps in both at once for better accuracy
View 1 Image
Plans call for the EarEcho system to work in either of the user's earbuds, or perhaps in both at once for better accuracy
1/1
Plans call for the EarEcho system to work in either earbud, or perhaps in both at once for better accuracy

Whether we like it or not, there are now a lot of people walking around with earbuds jammed into their ears. Scientists from the University at Buffalo decided to see if they could put those buds to extra use, so they designed a system that allows the devices to help verify smartphone users' identities.

Known as EarEcho, the technology was created by a team led by Assoc. Prof. Zhanpeng Jin.

It utilizes modified commercially-available wireless earbuds, which are paired with a smartphone. One of those buds emits an audio signal – which could be any sort of sound – that is both absorbed and reflected by the inside surface of the wearer's ear canal. A small microphone that has been added to the bud picks up the reflected signal, and transmits it by Bluetooth to an app on the phone.

Given that everyone's ear canal has a unique geometry, the manner in which the audio signal gets absorbed and reflected is specific to each person. Jin tells us that by comparing the currently-reflected signal to one that was initially recorded as a baseline for the phone's registered user, the app can therefore determine whether or not the person who's presently wearing the earbuds is indeed that user.

A diagram of the EarEcho system
A diagram of the EarEcho system

The U Buffalo team tested the EarEcho technology on 20 volunteers, utilizing audio signals that consisted of speech, music or other content. These tests were carried out in environments such as a street and a shopping mall, with the participants both standing and sitting. Acoustic signal-processing techniques were used to limit interference from ambient noise.

The system proved to be about 95-percent accurate at authenticating identities based on a 1-second ear-monitoring period, although that figure rose to 97.5 percent when the period was increased to three seconds. Once developed further, it is hoped that the technology could be utilized as an unobtrusive alternative to security measures such as the entering of passcodes, or the use of fingerprint scanners.

"Think about that," says Jin. "Just by wearing the earphones, which many people already do, you wouldn’t have to do anything to unlock your phone."

A paper on the research was recently published in the journal Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies.

Source: University at Buffalo

0 comments
There are no comments. Be the first!