New technology identifies even most distorted fingerprints

The three researchers who developed the technology have founded a Warwick Warp, a company that will market it for use with access control systems, financial-transaction systems and passports.

U.K. researchers have developed a technique for identifying fingerprints that have been smudged, clipped, distorted or otherwise badly recorded. The researchers, from the University of Warwick, also claim the identification only takes seconds to execute, no matter the size of the database the prints are being compared against.

"Our technology ? provides high speed and more importantly, our system's accuracy and speed doesn't degrade when the size of database increases," said Li Wang, one of the researchers on the project, in a statement.

The technique involves building a master coordinate map for all the fingerprints on record, called an image space. When a smudged, uneven or distorted copy of a fingerprint needs to be identified, it then can be overlaid onto this topological map, allowing the system to make a match with the copy on record.

This approach is quicker than the fingerprint matching technique now deployed, which usually involves identifying a few key features on each fingerprint and matching those against a database of templates, one by one.

This approach is also the first that uses sweat pores as comparison points, which current systems can't do, according to the researchers.

The three Warwick researchers, Wang, Roland Wilson and Abhir Bhalerao, have founded a company, called Warwick Warp, to commercialize the technology. They plan to market the technology for use with access control systems, financial-transaction systems and passports.

The researchers tested the technology last summer at the London Science Museum, where they obtained the fingerprints of 500 visitors, including many children who tended to smudge their fingerprints.

"Children often tend to twist their finger when placing the finger on the scanner, creating an elastic deformation which provides a great testing ground for our technology," Wang said. Despite the antsiness of the children, the system achieved 100 percent accuracy in identifying the prints, Wang said.

Joab Jackson writes for Government Computer News, 1105 Government Information Group publication.

NEXT STORY: Input predicts AF solutions focus