For more information, contact firstname.lastname@example.org or 978-934-3224
Fingerprinting is arguably the most common and universal method of identifying people. This technique has been widely adopted since the early 1900s, when Scotland Yard and the New York City Police Department used it to identify suspected criminals.
The standard way of obtaining fingerprints involves rolling the fingertips on an inked pad and pressing them on a white card to create well-defined impressions. The fingers’ pattern of dermal ridges and grooves are then compared with prints on file to authenticate the person’s identity. Today, the messy ink pad is being replaced with a flatbed scanner, which scans the finger pressed against a glass plate and downloads the resulting image to a computer for comparison with a database.
An inherent problem with both techniques is that the fingerprints become distorted whenever the fingers are pressed against the card or glass. “Pressing the fingers can change the spacing between ridges by more than 20 percent,” says Prof. Sam Mil’shtein, director of UML’s Advanced Electronic Technology Center. “This can affect the accuracy of the identification.”
Mil’shtein and his students have developed a novel contactless fingerprint scanner that can produce line scans of fingers without any distortion. The scanner uses a high-resolution electronic camera that swings 180 degrees around a finger to record a three-dimensional, distortion-free scan. As an added security measure, the same camera also takes a simultaneous infrared image of the finger’s blood vessels. Like the blood vessels in the retina, a person’s pattern of finger blood vessels is unique to each individual, according to Mil’shtein. The entire scanning process takes only about a second.
This technology has many military, civilian and commercial applications. “The most obvious are in the fields of law enforcement, forensic investigation and counterterrorism,” he says. “For example, the scanner can be used not only to ID people at airports and border crossing checkpoints, but also to screen employees and job applicants, control access to secure areas in schools, offices and warehouses, confirm the identities of missing persons and many more.”
Mil’shtein says that through special algorithmic processing of the images, his group was able to achieve 100 percent success rate in matching full fingerprints, that is, prints showing the “core” of the finger. With partial prints, their success rate was 95 percent. “The system is highly accurate and reliable,” he says. “Current technology requires the core to match prints. With our device, partial prints are not a problem.”
Mil’shtein and his team presented their working prototype at the IEEE International Conference on Technologies for Homeland Security held May 12ߝ13 in Waltham. The meeting was attended by representatives from the Department of Homeland Security and the state government as well as from industry and business communities.
“The response at the conference was great,” says Michael Baier, an electrical and computer engineering sophomore from Medway, who led the team in developing and testing the system’s control electronics. ECE graduate students Anup Pillai and Ameya Shendye helped develop the algorithm. “We’re now looking into the commercialization of the technology,” says Baier.
So far, Mil’shtein has spent about $30,000 of grant money to build the prototype. He plans to conduct a test run at Boston’s Logan airport this summer. He is also working with an ECE junior, 23-year-old David Mailloux of Methuen, in applying line-scanning and algorithmic processing to an automated facial recognition system that can scan the bone structure of a human face. “Such a system cannot easily be fooled with simple disguises or makeups,” says Mailloux. “A person would need major plastic surgery of the eyebrow ridges, cheek bones and chin to render his or her face unrecognizable to our system.”