December 2001

Recognizing the Enemy

Continued from page 2

By Alexandra Stikeman

Even before the September 11 attacks, the U.S. Defense Advanced Research Projects Agency (DARPA) had begun its own research efforts in face recognition and other biometric technologies. Indeed, face recognition is a key part of DARPA's four-year, $50 million "Human ID at a Distance" project, launched in February 2000 to develop a sophisticated surveillance system for U.S. embassies and facilities at home and abroad (see "Big Brother Logs On," TR September 2001). The system would combine cameras, radar and various types of sensors to identify people up to 150 meters away not only by their faces, but also by the way they walk, for example, or measurements like girth and leg length. The project not only illustrates the government's intense interest in using biometrics to enhance national security, but could also serve as a window into the highly sophisticated systems that might be used in the future.

As promising as face recognition and other biometric technologies look, though, one overriding caveat remains. Simply put, to use biometrics to catch bad guys, you have to start with some idea of who the bad guys are. Surveillance cameras could check each face in an airport against the FBI's database, for example, but if the agency hadn't turned up the fact that the man at the coffee stand in the blue coat had trained at an al-Qaeda camp, and his photograph wasn't in its database, the check would have done no good. It's still unclear, in fact, whether the FBI had Atta's picture on file before the September 11 attacks. In other words, biometric data would have to be supported by good intelligence to be a powerful weapon in the battle against terrorism.