Law Enforcement Turns to Face-Recognition Technology

Article excerpt

When the 1950s television hit Candid Camera took America by storm, the outlandish situations captivated viewers in living rooms all over the U.S. A hidden camera captured the victim's reactions, and at some point, the prank was revealed with the signature phrase, "Smile, you're on Candid Camera!"

In the entertainment world, capturing someone's face on camera has continued to evolve. The intrigue is less about the situational story line and more about technological prowess. From Mission: Impossible and Enemy of the State to television shows such as CSI, 24, and Las Vegas, Hollywood generates story lines that often feature face-recognition technologies. The reality is that face-recognition technology--a form of biometrics--is a rapidly growing tool for real-world issues including public safety and security.

Biometrics and Face Recognition

The term "biometric" is a combination of "biology" and "measurement," one of several technologies used to identify individuals by the body's physical characteristics, along with fingerprint readers, iris scanners, voice analyzers, and computer-linked cameras that recognize the way people walk. Law enforcement has been learning how to use face recognition and other biometrics for public safety and information security. This controversial technology also raises ethical questions and concerns about law enforcement's use of face recognition in criminal investigations as well as covert operations.

How It Works

A face-recognition biometrics system analyzes the characteristics of a person's face, which is usually captured by a digital video camera. The overall facial structure is measured, including the distance between the eyes, nose, mouth, and jaw edges. These measurements are saved in a database and used for later comparison, if needed.

Face-recognition technology can also be used for surveillance purposes. Unlike iris scans, hand geometry, fingerprint recognition, or voice recognition, face recognition does not require close proximity to the donor for identification. So, face-recognition technology can be used without the subject ever knowing he or she is the focus of such a scan, which opens the door particularly in intelligence and covert operations.

The distinctions for using face-recognition technology are subtle yet compelling. Law enforcement must decide whether it will use the technology for verification versus identification. Verification implies face-recognition biometrics that will be used to answer the question, "Is this that person?" Identification, on the other hand, implies that the technology will be used to answer the question, "Who is this?" by reading a facial sample and comparing that sample against a database.

Historical Perspective

Face-recognition biometrics actually dates from the late 19th century's Bertillon measurement system. In the 1880s, Parisian anthropologist Alphonse Bertillon pioneered the system based on a belief that a person's bone structure remains unchanged after age 20. Bertillon focused much of his work on the shape and breadth of the head and face. Suspected criminals were relegated to physical exams that included having their heads and bodies measured, recorded, and manually compared to Bertillon's elaborate cataloging system.

The manual record filing and checking system was quite fast for its time. Bertillon's system spread worldwide for nearly 2 decades until the system was derailed by inconsistent measurements of the same people, resulting in misidentifications and the jailing of the innocent. The Bertillon system was summarily discredited and abandoned in favor of fingerprinting.

Then, computer-vision technology was born in the late 1950s, using rudimentary programming to detect images. Through the 1960s and 1970s, experts increasingly experimented with new ways to enhance the computer's ability to distinguish one image from another, eventually spawning the growth of face-recognition biometrics. …