Prepared Remarks by Arden L. Bement, Jr.
Before I focus on our biometrics efforts, I would like to give you a brief overview of NIST's overall mission. As an agency of the Commerce Department's Technology Administration, NIST works with industry to develop measurements, standards and a variety of technologies. Our efforts are designed to enhance American productivity, facilitate trade and improve the quality of life.
Established in 1901, NIST is the only federal agency that focuses so intensively on standards and technology. During the past century NIST researchers have helped to pave the way for the innovation and robust economic growth that have made the United States the world's most prosperous nation.
Technological innovation underpins sustained periods of economic growth and solid increases in productivity that ultimately benefits everyone. We have a staff of some 3,000 scientists, engineers and other personnel, and about 1,600 visiting researchers. Our current annual budget is about $819 million.
We have four related programs: Our laboratories, including the Information Technology Laboratory that works with the biometrics industry. The Baldrige National Quality Program, which promotes excellence in business, health care and education. The Advanced Technology Program, which funds high-risk private sector research on promising technologies that have the potential for making a broad impact on economic growth. And the Manufacturing Extension Partnership, in which NIST works with 2,000 manufacturing specialists and staff at affiliated centers around the country.
Last year's terrorist attacks on the Pentagon and the World Trade Center taught us a great deal about our strengths and weaknesses as a nation. We witnessed great courage on the part of men and women in the military, fire fighters, police officers and ordinary citizens. And many of us experienced an unfamiliar sense of vulnerability. Yet, in some ways, the attacks have backfired on the perpetrators. They sparked a sense of solidarity among Americans, and strengthened our determination to enhance the security of our citizens.
President Bush, the entire administration, and the Commerce Department are committed to strengthening homeland security while maintaining American leadership in science and technology and accelerating the pace of scientific discovery and technological innovation. As director of NIST, I believe that a strong technology base will play a key role in helping the United States and its allies win the war on terrorism. One way we do this is through cost-shared funding of private sector research projects. Another way is by organizing consortia, some involving both corporations and universities, to work on specific projects.
Yet another way is via voluntary standards development, working with standards development organizations at home and internationally. One of our most important goals is to ensure that U.S. technologies and U.S. companies are not hampered in world markets by ill-considered or restrictive standards. And we want to make sure that different companies products and services are interoperable.
In the biometrics arena, NIST has worked with industry for years. In 1999 we played a significant role in the unification efforts of two industrial groups developing biometric application programming interfaces. The development of a single approach assured biometrics companies and their potential customers that different biometric devices could exchange information efficiently.
In our view, biometric technologies are becoming a needed foundation for secure identification.Biometric technologies can support homeland security, prevent ID fraud and play a role in supporting confidential financial transactions. We can use them to secure everything from a military installation to a school lunch program.
Of course, NIST has been involved for decades in working with the original biometric: the fingerprint. NIST worked closely with the law enforcement community to develop the first automated fingerprint matching system. For more than 30 years we have worked with the Federal Bureau of Investigation to automate the process of scanning their fingerprint files. The FBI has more than 35 million fingerprint cards. Computerized matching of fingerprints helps law enforcement officers in the field and supports forensic research.
Just a few years ago NIST computer scientists did some innovative work that significantly extends the range of fingerprint matching capabilities available to law enforcement officers. Working with the FBI, we developed software that enhances low-quality fingerprints for electronic matching.Low-quality fingerprints are precisely the kind you are most likely to find at a crime scene. They are the opposite of the carefully done prints you get when a suspect is booked at a police station. Those are relatively easy to match electronically. Trying to make a match based on the latent fingerprints found at crime scenes is much more difficult. Typically, investigators have to work with smudged, partial prints that are naturally of poor quality.
Until recently, matching these crime scene fingerprints electronically with those in the FBI's database was almost impossible. The new software we developed in cooperation with the FBI makes it possible to search the entire FBI database, instead of only part of it. It also allows law enforcement agencies in different locations to exchange fingerprint information directly, instead of always working through a national database. The software speeds up and automates what had been a very laborious process.
NIST is spearheading the Face Recognition Vendor Test 2002, which is evaluating automated facial recognition systems that eventually could be used in the identification and verification process for people who apply for visas to visit the United States. The significance of the Face Recognition Vendor Test 2002 is evident by its large number of sponsors and supporters; this includes sixteen government departments and agencies. The current evaluation builds on the success NIST personnel have had in evaluating face recognition systems over the last decade. The evaluation methodology developed for FRVT 2002 will become a standard for evaluating other biometric technology. We will learn precisely how accurate and reliable these new systems are.
Fourteen companies participated in FRVT 2002, with the result to be released publicly in November 2002. We deliberately designed a tough test that involved matching extremely challenging real world images. It required participants to process a set of about 121,000 images, and match all possible pairs of images from the 121,000 image set. In other words, this required some 15 billion matches. As you can imagine, this generated a mountain of data, and we are crunching all the numbers to see how well the systems worked. At the Biometric Consortium Conference we had two presentations on FRVT 2002. Yesterday, Jonathon Phillips gave and overview the FRVT 2002, and tomorrow Patrick Grother will give some more details on how we use evaluation statistics to measure the performance of biometric technologies.
Our goal is to give government agencies a reasonable expectation of the capabilities and limitations of these facial recognition systems. We did this work to meet the requirements of the Patriot Act, which Congress passed in October of last year. We want to thank all of you who came forward to participate in these demanding tests. And we want to thank all the federal agencies that have worked with us on this project, especially DARPA and the U.S. Navy, which allowed us to use its Dahlgren facility during the testing. Ultimately we will make a recommendation about which biometrics system, or combination of systems, would best be used to secure the nation's borders. We expect to issue a report on our findings in November.
NIST is happy to be a partner in the biometric community, supporting both industry and government users. We co-chair the Biometric Consortium and its working groups. In addition, NIST co-sponsors Biometric Consortium conferences such as this, and sponsors biometrics technical development. NIST participates in the Consortium and in many government groups. We recognize that open consensus standards for biometrics, and associated testing, are critical to providing higher levels of security through automatic personal identification systems. These technologies can also play an important role in preventing identity theft.
In the past year NIST has worked in partnership with U.S. industry and other federal agencies to establish formal groups for accelerating national and international biometric standardization.The executive board of INCITS (International Committee for Information Technology Standards) established its Technical Committee M1 on Biometrics in November of last year. In June, the ISO, which is the international standards body, established a new subcommittee on biometrics. These new efforts will provide the needed venues for a focused and comprehensive approach to the rapid development and approval of formal national and international biometric standards. They will include developing standards for common biometric data structures, publishing standard reference data collections and developing interoperable architectures. NIST is pleased to be able to offer a highly qualified individual to serve as chairman of these new bodies, Fernando L. Podio from the Convergent Information Systems Division of NIST's Information Technology Laboratory, the Co-Chair of the Biometric Consortium. These are major roles, and will take considerable time and effort.
Bridging from the national to the international work will not be an easy task. Nevertheless, we consider this effort to be an excellent opportunity to accelerate the deployment of standards-based biometric technologies.
Both the national and international communities need this work to be done, and time is a compelling factor for new homeland security applications. As you can see, there is much important work still to be done. Yet we have limited resources. We must find innovative ways to work together.
When the private sector and universities team up with federal and state government, we succeed in leveraging the available resources. The Biometric Consortium is a good example of how this can be done. That is why I am delighted to see so many of you participating today. I hope you already have had a very successful conference and that you all will leave here having made important progress on what clearly is a pressing national and international challenge.