Your Smartphone Could Soon Recognize You By Your Unique Hand Vibrations

First, there was fingerprint identification, then there was face recognition, and now it seems like your phone could know it’s you just by the touch of your hand alone. 

Developed at the Toulouse Institute of Computer Science Research in France, an artificial-intelligence system was taught through deep learning how to pick up on vibrations in a person’s hands and use them as identifiers when holding a phone.   

The system known as ‘HoldPass’ uses a heart activity biometric trait that reads your cardiac cycle, also known as ballistocardiography (BCG). 

While BCGs have been used as biometric authentication before, using a hand (Hand-BCG) to access the cardiac cycle has not been done before. Needless to say, neither has to put that technology into a phone been accomplished. 

Over 200 people were invited to take part in the in-depth study over a period of several months to get their heart readings right. 

While there were some worries about things like motion and weak signals hindering such technology, HoldPass bypassed all of these by presenting an alignment-free authentication program. 

The volunteers were studied at great lengths and a dataset of over 1200 readings was collected to implement into the system. With this, HoldPass was then able to accurately learn these readings.   

According to the study, “multiple feature candidates” were implemented to better help the system recognize each individual user. The team has also introduced a ‘Cycle Alignment Error’ to mitigate the complexities of reading heart activity-based signals.   

All of this can then be translated into the standard sensors already found on smartphones.   

[via New Scientist and Toulouse Institute of Computer Science Research, cover image via tonktiti/Adobe Stock]