Some Samsung weeks ago it gave much to talk about when moving into the background the use of readers of fingerprints as a means of authentication of identity to their smartphones, giving preference to the facial recognition and iris scanning. It now appears that it was indeed movement wise and visionary, because that digital technology would not be so sure.
It turns out that the New York University (NYU), in conjunction with the Tandon School of Engineering and the College of engineering at Michigan State University have published the results of a new research on the vulnerabilities of systems of reading fingerprints on smartphones, discovering that this security measure could be ignored without much complication with a sort of “master fingerprint”.
The project came from the idea of creating this master mark, finding more matching authentication points in a sample of 8,200 partial records of fingers of different users. At the beginning per 800 partial impressions, found 92 master tracks potentials that could violate 4% of crashes.
But from this the researchers decided to create an algorithm to unify different synthetic models of master tracks, composed by different points of identification of those 92 potential assemblies.
As I was feared, to run tests with these artificial tracks readers could be deceived more than in the first stage, believing that the registered fingerprint pattern actually belonged to the rightful owner.
With a range of variant effectiveness between 26 and 65% of cases, depending on the model of master mark applied, this research checks that those readers may not be so sure.
Follow us and comment on Facebook.
