Biometric trends in security
The biometric era that many foresee should be based on several biometric identifiers, rather than a single trait, a prominent biometric researcher contends.
Government and commercial interest in biometric technology, which measures unique body characteristics such as fingerprints or iris patterns to determine identity, has swelled during the last several years as the vulnerabilities of traditional security measures have become increasingly apparent. While no single biometric identifier produces completely accurate recognition results, a combination of biometric measures could give system users nearly perfect accuracy, said Anil K. Jain, a Michigan State University computer science and engineering professor who has written several books on biometrics.
“One way to improve accuracy is to combine biometric traits,” Jain said. “I think this is the way we should go.”
Jain discussed what he described as the “new paradigm for security” at a Jan. 25 seminar sponsored by Mizzou Engineering’s Center for Cyber Security Research. The year-old center focuses on such computer system security issues as biometrics and sensor network security.
Biometric systems are already at work in U.S. military and business operations in the form of fingerprint sensors, but such scanners can be fooled, Jain said. He said that adding face and voice recognition biometrics—which would require no more technology than that already available on many cell phones—could greatly reduce the fingerprint identifier’s error rate.
Researchers are investigating other biometric characteristics, including ear shape and gait patterns, in hopes of coming up with new measures that would be both unique and permanent as well as universal, he said. Jain believes a combination of these new and older methods can foil attacks as well as decrease identification errors.