DURHAM, N.H.—The University of New Hampshire and Spaceflight Systems Corporation of Bedford have won a grant from the U.S. Army Research Office to test new biometrics technology that could improve homeland security while reducing its cost.

UNH electrical and computer engineering Professor Andrzej Rucinski has partnered with the company to develop a single piece of hardware to analyze more than one kind of biometric data—fingerprints and face recognition, for example. The U.S. Army Research Office is supporting the work with a $100,000 Small Business Technology Transfer (STTR) grant. If the technology proves itself, the Army could provide up to an additional $1 million to commercialize it. The N.H. Industrial Research Council is also supporting the research with a $30,000 grant.

A “single engine” does more than save money by having one machine do the work of two. It can also increase performance. Instead simply giving an answer for each biometric characteristic evaluated, which can be conflicting, the device can fuse them to give a single response.

“What we’re trying to do is use fingerprint and face together to positively identify humans,” explained Dwayne Jeffrey, principal technical staff at Spaceflight. “If you take just fingerprint or just face alone, sometimes it’s easier to fool the system.”

But the device would not need to have both types of data to function. “In real life situations, sometimes you don’t have a terrorist’s fingerprint,” said Rucinski.

Looking at more than one identifying characteristic should cut down on cases of mistaken identity. “No single method is 100 percent right. “We need multiple ways to recognize someone,” said Rucinski. The worst possible scenario is a false positive, because you can ruin the life of a person.”

Spaceflight is planning to spinout a separate company, called SEA Biometrics, to commercialize civilian applications of the technology. The acronym, chosen by Rucinski, stands for Single Engine Adaptive.

The technology has thousands of potential applications, according to Rucinski. “This could be used in ATMs, cars, computers, security area protection,” he said. “It recognizes patterns, so it could also have applications beyond identifying people, such as chemical analysis, signature recognition or analyzing mammograms for better breast cancer diagnoses.”

###