Biometric Hell: Police Across The US To Use Face Scanners To ID Suspects

07-25-2011 • singularityhub.com

The gizmo, called MORIS–Mobile Offender Recognition and Information System–has a built in iris scanner and biometrics analysis software. The officer holds the scanner about 5 or 6 inches away and it automatically detects the iris and takes a high-resolution image. Like other iris scanners, the MORIS system identifies 235 distinctive features in each iris. It’s like a fingerprint for the eye–assumedly no two are alike. An algorithm is then used to search for a match with the signatures of others in a database. Pictures of faces can also be used. The picture is snapped from about 2 to 5 feet away and about 130 unique features on the face are identified and metrics are recorded such as the distance between the eyes. Again the database is accessed to see if there’s a match. MORIS is an offshoot of a technology called IRIS. Developed by BI2 Technologies, IRIS–Inmate Identification and Recognition System–was initially used to address the rather unsettling problem of mistaken inmate identity (last year a Rhode Island inmate escaped by assuming the identity of another scheduled for release). More than 320 law-enforcement agencies in 47 states are using IRIS to keep track of their inmates. The MORIS device combines iris and face recognition with the age-old method of offender identification: fingerprinting. There’s a scanner pad on the device, but when police officers in Brockton, Mass. tested a MORIS prototype the fingerscanning feature performed poorly. Regardless, the convenience and accuracy of the iris and face scanning functions makes MORIS a must-have realtime tool for cops on the go. Instead of calling in license plate numbers or relying on questionable ID numbers cops get quick and highly-accurate answers. And the device is a godsend for identifying people without IDs such as accident victims or homeless people. A physics lab in the UK tested the accuracy of BI2’s iris scanning technology. In over two million samples there wasn’t a single false match.