As we all know, humans are not very good at creating secure passwords.

To this end, technicians have invented a series of new devices that rely on biometric authentication (such as fingerprint recognition, speech recognition and face scanning) to solve (or to some extent) this problem.

However, recently, researchers at New York University and Michigan State University questioned whether biometric technology is safe enough.

Phillip Bontrager, a doctoral student at New York University and principal researcher of the study,

“Fingerprint based authentication is still an effective way to protect devices or systems, but most systems do not verify whether fingerprints or other biometric features are from real people or copies.”

The problem lies in the way most fingerprint sensors work.

Previous studies by Nasir Memon, a professor at New York University, have described in detail some fatal defects in the system. For example, most of these systems rely on partial fingerprints to confirm identity, rather than full fingerprints. To this end, most devices allow users to submit multiple fingerprint images, which are usually sufficient to confirm their identity as long as they match any saved part.

Recently, researchers have used these findings to create a new machine learning algorithm to generate the so-called “fingerprint” by collecting the fingerprint images stored in the fingerprint accessible system. In the future, some people may use these fake synthetic products to attack the real system or verification device, test all fingerprints in the system library, and finally open the device.

At present, the system has not been tested on real equipment.

We believe that biometric devices can protect our increasing sensitive data, so we think this discovery is also very important.

Leave a Reply

Your email address will not be published. Required fields are marked *