Researchers at New York University (NYU) and Michigan State University (MSU) are attempting to develop a revolutionary new means of bypassing biometric fingerprint authentication. To do this, developers are attempting to create synthetic digital fingerprints compromised of multiple images of actual fingerprints superimposed onto one another. Using these different combinations of fingerprints, researchers then attempted to employ a dictionary-style hacking attack against biometric authentication systems, hoping to fool, trick or bypass them. What they found was interesting.
Within a given margin of error, hackers were able to successfully bypass every biometric fingerprint authentication system in front of them. More specifically, hackers were able to bypass systems with a 1% False Match Rate (FMR) 76.67% of the time. At a FMR of 0.1% hackers were able to bypass the system 22.5% of the time and at a FMR of 0.01% hackers were able to bypass the system 1.11% of the time. Their research paper, entitled “DeepMasterPrints: Generating MasterPrints for Dictionary Attacks via Latent Variable Evolution,” outlining their design, methodology, analysis and conclusions, was officially released to the public earlier last month.
See full pdf below, or download for yourself here: https://arxiv.org/pdf/1705.07386.pdf