Big Brother Greets Visionics

0 Comments | Insight on the News, Oct 22, 2001 | by Sheila R. Cherry

New face-scanning technology promises police the ability to find and track criminals using surveillance cameras, but critics charge that its use tramples individual rights.

Terror attacks on the World Trade Center and the Pentagon on Sept. 11 have some civil libertarians rethinking their privacy concerns and aversions to such surveillance technology as face-scanning software. They are asking, if privacy is sacrificed to technology for the sake of security, will it be effective?

According to Kevin Watson of the Law Enforcement Alliance of America (LEAA), a coalition of law-enforcement officials, crime victims and concerned citizens, face-recognition technology requires a large database of known suspects. That terrorists were able to board airliners that they later used as flying bombs was not just a matter of unknown criminals slipping past sleeping authorities, Watson points out. "They even used their own names."

Watson cites a long list of steps authorities should take to combat terrorism before spending law-enforcement resources on more technology. Security-detection equipment is only as good as the people operating and supporting it, he notes, so he advocates replacing undertrained, minimum-wage security employees with professional law-enforcement personnel.

Watson seethes that Visionics Inc., the Minnesota- and New Jersey-based biometrics company that manufactures face-scanning technology, began an intense promotional campaign of its products immediately after the terrorist attacks. "These guys are looking to make a buck," he says. "I personally think it was shameless, particularly when they don't have a proven track record of effectiveness."

What is this new technology? Joseph J. Atick, cofounder and chief executive officer of Visionics, has invented a process that allows computers to recognize faces in much the same way that humans do. The key to his face-scanning technology, Atick explains to Insight, is that intelligence organizations worldwide now have images of people with known terrorist connections they can share with each other to create a reliable "watch list" that can be stored in databases and used to match faces in a monitored crowd. He describes it as a shield against terrorism.

"Biometrics" basically are technologies that establish the identity of a person based on unique physical characteristics. It could be a configuration of a face, voice, fingerprint or any other physical characteristic or combination of characteristics. Atick, the former director of the Computational Neuroscience Laboratory at Rockefeller University, was one of the inventors of biometric technology. He focused on facial recognition because he thought it had the broadest range of applications.

Atick discovered the algorithm that allows computers to do "pattern matching," which characterizes and records patterns of the human body. It measures deviations, he says, and creates a map of the face so that it can identify a person analytically. The facial-recognition software, FaceIt, is one of the newest products in the inventory of Visionics, which prior to Sept. 11 was estimating year-end earnings of more than $30 million. The stock has since tripled.

Software that tracks individuals without the need for human recognition now has governmental agencies, banking institutions and airports clamoring to Visionics for help in securing access control and thwarting identity fraud. Before the attacks, more than two dozen law-enforcement organizations had added FaceIt to their technology toolboxes, including police forces in New Jersey, California, Texas, Minnesota and Oklahoma.

Most police customers use their facial-recognition systems when booking those arrested to check the suspect's face against the department's database of mug shots. The level of recognition is variable and can be calibrated from a low-percentage match or resemblance to a high-percentage or exact match against stored images. When a match is made between the suspect and a face already in the database, the software freezes the image and stores it for an officer to review.

But the recent decision in Tampa, Fla., to use FaceIt as a crowd-scanning alarm to identify criminals outside of the controlled environment of a police station brought howls of public outrage. The on-the-street face-identification cameras were the first of their kind to be used within the United States. Then public fears were further inflamed when a picture used to illustrate a display of the cameras in a recent news story led to a much-publicized misidentification (by humans, not the technology, Atick bristles).

The crowd-scanning alarm systems are used in three other cities: London and Birmingham, England, and at Iceland's Keflavik Airport. Bush-administration officials are considering its application for U.S. airports. The crowd-monitoring version of FaceIt sounds an alarm when it matches the face of an individual in a crowd with someone who has a certain kind of criminal record or who is on a watch list.