This article is from the archives of the UB Reporter.
Archives

CUBS focus is biometrics

Published: January 20, 2005

By ELLEN GOLDBAUM
Contributing Editor

Biometrics, the science of identifying individuals based on their physical, chemical or behavioral characteristics, is a key piece in homeland security strategies, but no single biometric—such as face, signature or fingerprint—fits all applications, says the director of UB's Center for Unified Biometrics and Sensors (CUBS).

Many of the systems on the market have a high false-positive rate, which could be misleading or even dangerous, according to Venu Govindaraju, director of CUBS and professor of computer science and engineering.

For that reason, he says, the UB center takes a unique approach to developing technologies in biometrics, combining and tuning different biometrics to fit specific applications.

"We believe success in this area depends on being able to combine and tune technologies to different applications by using contextual knowledge about how the data will be used," said Govindaraju. "The technology for these applications exists. Now it's a question of figuring out how to build the best devices."

CUBS has secured funding to create systems that:

  • Could quickly identify suspicious public-health patterns, indicating a possible terrorist attack or epidemic, by automating handwritten data collected about patients entering the nation's emergency rooms. Funded by the National Science Foundation.

  • Will help close the digital divide between the English and non-English speaking worlds by developing the first optical character reader (OCR) software for handwritten and machine-printed Arabic documents so that they can be searched automatically with keywords. Funded by the Director of Central Intelligence Postdoctoral Research Fellowship Program. (See story on this page for further details.)

  • Protect Web sites against cyberterrorism using handwritten CAPTCHAS (Completely Automated Public Turing Tests to Tell Computers and Humans Apart), automated tests or puzzles designed to determine whether visitors to sites are humans or machines. Funded by the Calspan-UB Research Center.

  • Bridge biometrics and forensics by matching automatically facial characteristics described by crime victims or witnesses with large facial databases of persons previously or allegedly involved in crimes or acts of terrorism. Funded by the Calspan-UB Research Center.

  • Prevent credit-card fraud and protect homeland security by combining face, fingerprint and signature biometrics, and embedding them in smart ID cards. Funded by the New York State Office of Science, Technology and Academic Research, UB and Ultra-Scan Corp.

  • Combine biometrics, such as signatures and fingerprints, to identify individuals entering the U.S. and to improve overall system accuracy. Funded by the U.S. Army Research Laboratory.

Additional funding for CUBS projects comes from other private high-tech firms, such as mobileLexis (Salt Lake City), Infinite Group (Rochester), Uniform Data System for Medical Rehabilitation (Amherst) and Ubique (Buffalo).