Facial recognition technology tested at the Palm Beach International Airport had a dismal failure rate, according to preliminary results from a pilot program at the facility.
The system failed to correctly identify airport employees 53 percent of the time, according to test data that was obtained by the American Civil Liberties Union under Florida's open records law.
"The preliminary results at the Palm Beach International Airport confirm that the use of facial recognition technology is simply ineffective and of no value," said Randall Marshall, legal director of the state ACLU chapter.
The manufacturer of the system, Visionics, said the results were poor because their product was not used correctly.
Ever since the Sept. 11 terrorist attacks, face scanning technology has been touted by manufacturers as the perfect device for recognizing terrorists in airports. In theory, the systems use surveillance cameras to scan crowds for bad guys and sound an alarm when a match is made between a live person and the system's database of known criminals.
The Palm Beach airport tried Visionics' FaceIt system, which snaps photographs of passersby using a security camera and breaks down their facial features into a numeric code that is matched against the photograph database.
The month-long test compared 15 employees against a database containing the mug shots of 250 airport workers, said airport spokeswoman Lisa De La Rionda, who declined to comment on the quality of the system.
"They never made promises to us about how successful the system would be," she said, stressing that it was tested free of charge.
But the ACLU said the study was done under optimal conditions and still exhibited fatal flaws. Out of 958 attempts to match the 15 test employees' faces to the database, the system succeeded only 455 times.
The Tampa police department has also been testing the FaceIt system over the last six months, and the technology has yet to make a match with a database of known criminals.
"The system could be serving as a deterrent for criminal activity ... we still believe in its potential for law enforcement," said police department spokeswoman Katie Hughes.
The airport trial found that the photographs included in the database had to be good quality to avoid false alarms and ensure successful matches. Head motion, indirect lighting, sunglasses and eyeglasses also flummoxed the system.
The finicky nature of the software was previously documented by Internet privacy and security consultant Richard Smith. Last fall, Smith analyzed the FaceIt software and found a 50 percent failure rate as he adjusted for variables such as face angle and hats.
"If you adjusted everything just right you could get OK results," he said.
Visionics, whose face scanning systems are being tested at four U.S. airports, bristled at the ACLU's conclusions.
"The decision makers will not be reading a report from the ACLU, they'll be looking at the real data," said Visionics spokesman Meir Kahtan.
He said that similar tests at the Dallas-Fort Worth and Boston Logan airports showed a 90 percent success rate and insisted that the poor results at the Palm Beach International airport were due to incorrect lighting. Results for the other pilot programs were not immediately available.
Scanning Tech a Blurry Picture
The Positive Side of Biometrics
Face Scanners Turn Lens on Selves
Hide Out Under a Security Blanket
