DeepMasterPrints: fingerprints that open everything! Researchers from New York University have created a set of master keys from fingerprints that can be used to breach biometric identification systems.
Effectiveness? The fingerprint database the researchers used had a chance of falsely matching a random fingerprint one in 1000 times, and the keys they generated matched one in five times.
Their study was published on ArXiv and demonstrates that fingerprints can be artificially created using machine learning and used to breach databases used by fingerprint authentication systems.
This is of course very worrying because more and more devices are using large-scale databases to uniquely identify users.
A study published last year by Counterpoint Research showed that over 50% of smartphones released in 2017 had fingerprint sensors and predicted that number would increase to 71% by the end of 2018.
The artificial prints, dubbed DeepMasterPrints by the researchers, can accurately mimic one in five fingerprints.
Using millions of fingerprints, neural networks created fake prints that look very convincingly like real fingerprints.
DeepMasterPrints can be used to breach a system that requires fingerprint authentication without even needing the user's fingerprints.
Read the study
https://arxiv.org/pdf/1705.07386.pdf
__________________
- Windows 10 Pro no more passwords: Palm-vein technology
- Prevent Windows 10 from syncing passwords between devices
- Hacking the most commonly used tools
