.
Researchers have shown it's possible to use a bone conduction speaker and microphone on Google Glass to accurately identity a user.
Researchers have shown it is possible to use a bone conduction speaker and microphone on Google Glass to accurately identity the wearer. The gadget is shown above
Computer scientists from the University of Stuttgart, University of Saarland and the Max Planck Institute for Informatics in Germany developed the prototype system dubbed SkullConduct,
The researchers said it was developed to boost security for the 'increasing number of eyewear computers'.
This is becoming a key area of research due to a new version of Google Glass expected to be released in the near future.
The researchers said secure user identification is particularly important for wearables that don't have keyboards, because PINs can't be used and there are limited ways to keep them secure.
Devices such as Google Glass are predominantly personal devices holding information about the wearer, including social media logins and even bank details.
The biometric system makes use of a microphone and bone conduction speaker built into an altered Google Glass device, The system analyses the frequency response of the sound after it travels through the user's skull to identify them. The process, which has a number of steps, is shown above
'Current systems, such as Google Glass, are not protected and attackers can simply put on the device and access sensitive information,' the experts said.
As a solution, they wrote in a paper: 'We present SkullConduct, a biometric system that uses bone conduction of sound through the user's skull as well as a microphone readily integrated into many of these devices, such as Google Glass.'
The system analyses the frequency response of the sound after it travels through the user's skull, to identify them.
It does this by sending a sound pattern into a user's skull – a type of white noise - and the records the noise that pings back.
This changed noise unlocks a device every time the action is repeated, so the Google Glass device 'knows' when it's being worn by a specific user.
As the study puts it: 'Individual differences in skull anatomy result in highly person-specific frequency responses that can be used as a biometric.'
The researchers explained bone conduction has only recently become available on eyewear computers, such as Google Glass, 'as a privacy-preserving means of relaying information to the user.'
The experts tested their prototype on 10 different people who were asked to take the device on and off to test the identification technique's reliability. They showed the frequency response was person specific and stable, even when the device was taken off (red) and worn again (blue) a number of times
It has however been used in different consumer devices such as hands-free headsets, headphones and hearing aids.
The experts tested their prototype on 10 different people.
They recorded nine men and one woman between the ages of 21 and 36 years in a quiet room where a fast blast of white noise was piped and recorded how the sound was relayed through their skulls, repeating the action 10 times each.
'After five recording trials, we asked participants to take off the device and put it back on to include different placements of the device on the participant's head,' they wrote.
In this way, they trained the system with multiple recordings from different people.
The team was able to show their device could identify users with 97 per cent accuracy.
It also showed the frequency response is person specific and stable, even when the device was taken off and worn again a number of times.
The trial suggests the technology could be used to identify owners of Google Glass, for example, and could be particularly handy if the same gadget is used by multiple users at work, for example.
But it may be a while before we are able to use our skulls as passwords in the real world, because the prototype did not need to cope with any background noise such as traffic and strangers talking, if it was being used on a street, for example.
No comments:
Post a Comment