Your next phone could unlock by reading your lips, ears, or even heart

A moment on the lips

Spoken passwords are all well and good until someone makes a recording of you speaking, or a talented impressionist impersonates you. One extra safeguard to make sure that the pass code speaker is really you? Watch a person’s lips as they speak it.

Researchers have explored a couple different ways to do this. At Hong Kong Baptist University, engineers have developed a computational learning model that examines the visual features of a person’s lips, including shape, texture, and movement.

Meanwhile, at Florida State University in Tallahassee, a similar goal is being pursued using sonar. Called VoiceGesture, the Florida State system effectively turns your smartphone into a Doppler radar, transmitting a high-frequency sound from the device’s speaker and then listening to the reflections on the microphone when a person says their password.

Both approaches can offer an extra level of protection on top of existing voice biometrics.

It’s not what you said, but how you said it

University of Michigan

Another possible solution to voice biometric spoofing is one pioneered by researchers at the University of Michigan. They’ve developed a wearable device — which currently takes the form of a necklace, ear buds, or glasses — and uses an accelerometer to measure the subtle skin vibrations in a person’s face, throat, or chest when they speak.

Finding a way of incorporating it into a consumer-level device may be tough, but it could certainly add another level of security to voice-based biometrics.

It’s all in your “heartprint”

You know those cheesy 1980s movies where a character is told that the way to resolve their angsty teen problem is to listen to their heart? Well, it turns out that the same is true for security — except substituting the words “angsty teen problem” with “biometrics spoofing attack.”

At the State University of New York at Buffalo, researchers have created a heart biometric system that uses Doppler radar (again) to find and receive information about the unique signature of a user’s heart motion and associated traits.

The system is reportedly as safe to use as any other Wi-Fi device, requires just 8 seconds of initial “heart data” to train, and — best of all — can be used for continuous authentication. That means that, rather than asking a user for their password just once when they log in, the system could continually watch to make sure it’s still you who is using it.

The sniff test

Pixabay / Creative Commons

“Smells like you. Come on in!” At least, that’s the basis for research from the Group of Biometrics, Biosignals and Security at the Polytechnical University of Madrid. Their work is based on the fact that everyone has his or her own odor, which remains steady, and this can lead to accurate identification of a person within a group that’s higher than 85 percent.

In theory, a bloodhound-style body odor “sniff test” to ID individuals could be a nonintrusive way to recognize individuals. In practice, a 15 percent error rate would be unacceptably high (compare it to Apple’s reported 1 in 1 million error rate for Face ID) and may be hard to gather in certain environmental conditions. This could turn out to be the basis for promising later research, though.

Butt-based biometrics

Unlocking your new smartphone using your unique butt print sounds like something that even Apple’s marketing wizards would struggle to sell to the masses. But that probably isn’t how this would be used. Pioneered by engineers at the Advanced Institute of Industrial Technology in Tokyo, the idea is to embed 360 sensors into car seats to analyze the, erm, size, shape, pressure points, and weight distribution of drivers’ posteriors.

Only those with the correct butt can start the car. That’s probably not good for heading back to work after gorging yourself over Christmas — but on the plus side, you could probably combine it with the smell sensor described above to make something hilariously gross.

Good vibrations

One issue with fingerprint sensors is that the actual “sensing” area they work with is typically limited to a small button on either the front or rear of a mobile device. That wouldn’t be the case with a similar finger-reading biometrics system developed by Rutgers University and the University of Alabama at Birmingham.

Their VibWrite system isn’t based on fingerprint recognition, but rather the unique signatures given off by finger vibrations. As a result of not needing fingerprint scanners, the inexpensive system could be applied to any surface — letting you add an authentication element to everything from opening your car door to switching on your desktop computer. Accuracy levels are currently in excess of 95 percent.

What’s going on ear?

What’s the one part of our body that’s virtually guaranteed to be pressed against our phones at some point? No, take your mind out of the gutter, we’re talking about our ears! The idea of ear-based biometrics has been explored for some time, with the idea being that each ear’s unique curves and other features could make for a brilliant fingerprint-style security system.

It’s an intriguing prospect, although having to imprint the side of your head on your smartphone every time you want to unlock it would be more than a little annoying. It would enable your device to continuously authenticate that you’re using the phone whenever you make a call, though!