New Method Discovered to Steal Biometric Data

A group of American and Chinese scientists have developed the PrintListener technique, which is a new method to capture a unique pattern of papillary fingerprint lines. This method is based on the analysis of the sound characteristics that arise during the sliding movement of the finger along the sensory screen.

Protection technologies using fingerprints today are extremely common and highly trusted. It is expected that by 2032, the market volume for fingerprint authentication will reach almost $100 billion. However, there is a growing realization that attackers can attempt to steal other people’s fingerprints.

According to test results, researchers claim that they can successfully attack up to 80% of partial and full prints in a framework of five attempts with maximum security settings. This is the first work using sliding sounds to obtain information about fingerprints.

According to the study, “the sounds of fingertips can be intercepted by attackers online with a high probability.” Attackers can access the microphone of the victim’s mobile device during the use of popular applications, such as Discord, Skype, Wechat, Facetime, and others, where swipes are used.

The developers of Printlistener faced a number of difficulties that they had to overcome. They were able to solve the problem of the weak sound of finger friction on the screen using spectral analysis methods. A statistical analysis was also conducted, and a heuristic algorithm was developed to filter the physiological features of a particular user.

Using the sound analysis technology, Printlistener scientists were able to restore the synthetic prints of PatternMasterprint. In realistic scenarios, this technique helped to successfully restore partial fingerprints in more than four cases, and complete prints in almost one out of ten cases. The results were better than the MasterPrint attack, which offers a random cluster of “universal” prints.

/Reports, release notes, official announcements.