UB scientists develop algorithm to assess security of partial fingerprints

Published 28 February 2006

Research is leading toward finding efficient methods of preventing unauthorized access to handheld devices, such as cell phones, wireless handheld devices and electronic audio players, as well as to secure Web sites, explains Venu Govindaraju, a University of Buffalo (UB) professor of computer science and engineering and director of the university’s Center for Unified Biometrics and Sensors (CUBS). It also will help make fingerprint matching for forensic applications more effective. The UB research addresses a key problem that has emerged in the quest for fingerprint access to electronic devices and Web sites: Quantifying how much security is possible with fingerprinting, given that most commercial sensors tend to capture only partial fingerprints. This problem needs to be overcome before it will be possible to routinely replace passwords with fingerprints,” Govindaraju said.

Called the Automated Partial Fingerprint Identification system, the algorithm developed by the UB scientists enables computer systems of, say, banks or online retailers, to determine whether or not to grant access, by securely matching two fingerprint images (the stored one and the “new” one) even when only part of the print is captured. The algorithm also takes into account the fact that even a legitimate fingerprint doesn’t always look the same due to the way a person presses on a pad or because of moisture or other factors.

-read more in this report