Facial Recognition Tests End in Failure

Sept. 5 -- In this week's Cybershake, we note how security officials are turning away from a security technology that received a prime amount of face time two years ago. Plus, we take a look at how to navigate the draw — and drawbacks — of online shopping.

Red-Faced Over Facial Technology?

Facial recognition systems made big news when it was installed on the streets of Tampa, Fla., in 2001. The technology allows for cameras to scan crowds of people and, using computer software, automatically matches faces against a picture database of known criminals.

And after the terrorist attacks of Sept. 11, 2001, airports such as Boston's Logan International, were interested in exploring if facial recognition could help screen against potential hijackers and trouble makers.

But two years later, it seems the promise of such systems is still nothing more than just promise.

Recently, government officials in Tampa removed the camera system from its streets, citing that it had failed to recognize any known criminals wanted by authorities there. A spokesman for the Tampa police department told The Associated Press, "It's just proven not to have any benefit for us."

The same is apparently true for authorities at the Logan International Airport in Boston.

"Back in January of 2002, we had a pilot program where we tested two separate facial recognition technologies at two separate checkpoints," says Jose Juves, a spokesman for the airport. The conclusion, he says: "It wasn't ready for primetime."

Juves and other airport officials say the system was just too costly and didn't seem to offer any significant advantages over less technical solutions.

"One of the things we found was that the technology itself was very labor intensive," says Juves. "It really involved someone sitting at a computer terminal and matching up the faces to the ones that appeared on the screen."

Logan airport officials also noted that law enforcement agencies lack a complete digital image database of known or suspect terrorists, further limiting the usefulness of facial recognition.