Facial identification is the fastest growing biometric technology today. According to many industry experts, it is also the most controversial of all biometrics. Despite their lingering questions regarding the practical usefulness of facial identification technology, law enforcement and military facial identification systems have been in place for several years without arousing too much controversy. According to industry insiders, this is because these applications have proven quite successful in carrying out specific objectives and the public is often unaware of these uses. Although facial recognition technology has not been proven to be an accurate and effective way of identifying terrorists or wanted suspects, some of the proposed post-September 11 uses of the technology – such as in immigration and airport security – have been welcomed by the general public. Are we too eager to buy into a new technology without clearly evaluating its effectiveness and without weighing the potential harms involved with its use?

Before September 11, the public viewed the technology with much more skepticism than it does today. Many people remain skeptical when it comes to widespread use of the technology in areas other than public safety and security. The possibility of identity theft and privacy infringement are the most common concerns voiced by people.

Much of the skepticism may be attributed to the very visible and troubling past uses of the technology. Tampa, Florida is one such questionable application.
Viisage Technology's software matched 19 faces in the Tampa Super Bowl crowd to persons whose purported likenesses were contained in a law enforcement database, all of whom had criminal records but were not wanted by police, nor were there outstanding warrants for the individuals' arrests. Ybor City, a nearby coastal village, installed 36 of the surveillance systems. Many visitors to the city and civil libertarians expressed their utter disgust with the system, calling it a "virtual lineup." The program was likened to Big Brother and people feared that the technology would worsen racial profiling practices and would accidentally identify innocent people as criminals. These criticisms persist even though the law is pretty clear, at least so far, that visual privacy rights don't apply in public places.

In spite of the harsh criticism of the Florida systems, Virginia Beach became the second city in the U.S. to use the technology when it installed a similar facial-recognition system in November 2001 at the Oceanfront. Although Mayor Meyera E. Oberndorf initially disapproved of Virginia Beach's plan to incorporate the technology for security purposes, she changed her opinion about the technology in light of the September 11 terrorist attacks. The public is more at ease with the idea of giving up a little privacy if it means the possibility of preventing even greater tragedies. "Possibility" is the keyword here, indeed. In actuality, the systems have yet to identify one "bad guy."

Some facial identification advocates say that the cameras have not caught any suspects because the systems have been successful in deterring terrorists and other criminals from entering the protected area. Others, like the single dissenter of the Virginia Beach decision, are more skeptical. Her research of the software led her to conclude that it was not worth giving up a sense of liberty for the marginal security benefits that these products provide.

For many, privacy is the most obvious and overriding concern. One aspect that bothers many people is both the incompleteness or the over-inclusive content of the database against which individuals' likeness is being compared. Virginia Beach police have minimized this concern with an agreement stating that they will only enter pictures of runaways, wanted felons and people suffering from dementia into the database. The Visionics' CEO announced that its system, FaceIt facial recognition, captures the images of faces taken from security cameras in airports and creates a unique mathematical identifier called a "faceprint" for each face. The faceprint is compared to those already stored in the database and if it doesn't match one of a terrorist, then the subject is not stopped and there is no record of that particular faceprint going through the system. The system simply doesn't recognize or identify the face.

Tampa police, who have a different system, have said that it is their practice to discard information on seemingly innocent faces in the crowd, but there is no requirement to do so. Nor is there any rule against the police scanning drivers license photos of passersby. The president of Visionics Corp. seems to be committed to minimizing privacy concerns involved with using his technology. He agrees that the federal government needs to regulate the use of facial identification systems.

Limiting the class of persons whose images are stored in a database is an important issue to be addressed by Congress. Although this may be a start, for many this doesn't safeguard the public from routine violations of privacy. Even facial recognition corporations admit the possibility of people getting unauthorized access to the data collected with the technology. The more widespread the technology becomes, the higher the privacy risks associated with its use.

Those concerned with privacy also point out the fact that the software only helps in situations where security services are aware of certain individuals. Some insist that these databases are not effective in catching unknown, foreign visitors to the United States and that those are the types of people we are trying to stop at the terminal.

The accuracy of the systems has also been a real concern. Even if a subject's face is stored in the database, a disguise or even a minor change in appearance, like wearing sunglasses or wearing or growing a mustache can often fool the system. Even an unusual facial expression can confuse the software. Facial identifiers often cannot distinguish twins. Other factors affecting the reliability of the images are changes in the lighting and the angle at which the photos are taken.

The industry is working hard to make the systems "smarter." After September 11, Visionics redoubled its efforts to create reliable facial recognition equipment. According to one National Institute of Justice Standards and Technology study, the industry still has a lot of work to do. The systems often have difficulty recognizing the effects of aging. NIJ's study found that digitally-compared photographs of people taken 18 months apart produced false rejections by the face identification program approximately 43 percent of the time. Manually controlled face identifiers, which are used by casinos, driver's license bureaus, welfare offices, for instance, are more successful than public surveillance systems because they allow the user to obtain high-quality photos of subjects. The pictures are taken in the same lighting, with the same camera, and the user typically requires the subject to look straight into the camera for the photo. These uses are not as troubling to those concerned with privacy infringement because in many of these circumstances, you give up a little privacy when you choose to participate in certain activities or services. The thought of being watched every time we step outside our homes, however, makes almost anyone squeamish, doesn't it?

Visionics' FaceIt technology measures 80 facial structures and calculates a "face print," which contains the physical measurements of your skull and your face. The company maintains that this face print is identity specific and is not affected by aging, viewing conditions, and superficial disguises. Other experts warn that the company, and other companies like it, may be misrepresenting the actual accuracy of their systems. While the actual technology is quite accurate, the systems that use the software are not always the best quality, nor are the databases' stored images. This is where the true problems with accuracy occur, according to industry leaders. Taking into account these factors, one company has estimated that the probability of capturing a terrorist, using facial recognition technology, is between 60 to 90 percent. Considering the deterrent effect this likely has on terrorist travelers, the use of these systems in airports may well be seen as a positive security measure.

Industry insiders rave about how exciting it is that this software is often compatible with the computers and cameras already in use by airports, banks, and other corporations. Because the technology is getting less expensive, it is becoming more and more attractive to private institutions as a way of preventing fraud. Many access-control uses of the technology are more acceptable when our safety and identity is in need of protection. However, public video surveillance uses of the technology raise unique privacy concerns that will need to be addressed in the courts.

The Senate is expected to address some very important issues regarding the use of these systems in airports, including whether or not the database information will be available on the Internet. Assuming that the law will require that database information be kept secure on discrete systems, the laws will also have to address how entities can ensure and maintain that level of privacy for the protection of their customers' or the public's rights.

The public understandably fears the worst with this modern and mysterious capability. There is no doubt that facial recognition systems can have a very positive impact on modern commerce as well as on crime-fighting, but there must be guidelines and limitations to the use of the technology to make it more privacy-friendly.

Some current and projected uses of the technology:

• The city of Santa Ana, California uses "MugMatch," a current generation automated mug shot system, which uses face pattern recognition algorithms, of which can be accessed and searched from a PC.

• In 1988, the Lakewood Division of the Los Angeles County Sheriff's Department started using composite drawings of a suspect or video images to conduct a search on its database of digitized mugshots. The same department also plans to search for suspects on a photo database of registered sex offenders

• . National Institute of Justice is funding projects involving the development and integration of software with facial identification modules. It has successfully integrated face recognition with still photographs. Still in the developmental stages is the integration of face recognition for video, newsgroup search agents, text categorization, and age progressio

• n. Gang Reporting Evaluation Tracking (GREAT) system can be queried using photographs of suspects so that law enforcement can circumvent any false identification cards and information presented by gang membe

• rs. In 1998, West Virginia became the first state to integrate facial identification technology into the drivers license application system. The system keeps a digital record of an applicant's image to prevent people from getting a driver's license under a false identity. Fingerprints can also be stored on the licenses to safeguard proper identification at stores and other locati

• ons. The Illinois State police and Secretary of State will have digitized the photographs of everyone licensed to drive in the state.
More than 100 American casinos use Biometrica Systems, a type of facial recognition system, to scan for cheaters and dishonest money counters

• . Integrated Law Enforcement Face-Identification System (ILEFIS) has been called the "next generation" face-identification technology. It incorporates a novel 3D-composite technology that can more easily identify angled-view face shots. This system will be helpful in correctly identifying non-cooperative suspects or subjects caught from a distance on a video surveillance camera.
The U.S. Army and West Virginia University have highly respected biometrics programs

• . West Virginia's Missing Children Clearinghouse uses the technology to look for missing children. Scanned pictures found on the internet are digitally compared to the images of children stored in its databas

• e. There is still talk about a national ID card even though such proposals have been rejected by the President's administration and the idea was not favored by most security industry experts. Other uses of biometrically encoded ID cards: passports, institutional and commercial uses for limiting access to secure areas or accounts, for verifying government benefit recipients, driver's licenses, and government ID car

• ds. The London borough Newham, numerous U.S. embassies, and the Los Angeles area have all used face recognition syst

• ems. Britain uses more than 200,000 video cameras for surveillance. Many of these are now being set up with facial recognition software. These systems generally use a computer that monitors the cameras, which are looking for known criminals. When the system identifies someone, the police are called to res

• pond. One new biometric system is the actual infrared heat pattern of the face. This allows the system to work in the dark.
Fresno International Airport uses facial recognition technology

• . The Israeli-Palestinian border uses both face recognition and hand geometry and will be incorporating INSPASS, a hand geometry system that is used for immigration in the U.S. and Canad

• a. The Department of Defense has a facial recognition system in place.

---

Return to Master Index