|News contact:||Steve McGregor, UTD, (972) 883-2293, [email protected]|
Four UTD Faculty Members Win $625,000
Part of Federal Government's Campaign to Combat Terrorism
RICHARDSON, Texas (April 16, 2003) - Four faculty members at The University of Texas at Dallas (UTD) have won federal funding of more than $625,000 for research efforts designed to improve face recognition technology, a potentially important tool in the government's war on terrorism.
The award was made by the United States Department of Defense and the Technical Support Working Group, the multi-agency federal task force whose goal is to develop and improve technology for combating terrorism. UTD's research proposal was one of less than 50 selected for funding by the agencies out of a total of some 12,500 proposals to strengthen the nation's security in the wake of the events of Sept. 11, 2001, according to university officials.
Face recognition technology is viewed by some as a possible boon to law enforcement because of the prospect of detecting and identifying criminals and other suspects through surveillance by video cameras and other means of crowds at airports, border crossings and other public venues. Through the use of cameras and sophisticated computer hardware and software, the face of an individual, either alone or in a crowd, can be compared with images stored in a database of persons deemed to be a threat and, in theory, matches can be made between the two. However, the technology has come under fire for several reasons, including its lack of reliability.
The UTD research to improve such systems will be conducted in two parallel efforts by faculty members in two of the university's schools:
Schweitzer and Truemper will apply new techniques they have developed, utilizing computational logic, which they believe will help compute facial features in images much faster than is currently possible. The pair's project will investigate how these techniques can be used to accelerate both the data collection and system run time, as well as its accuracy.
O'Toole and Abdi will create a means of evaluating the performance of a wide variety of face recognition systems against the performance of human subjects under similar conditions. The pair will undertake a series of experiments to evaluate how humans match the identities of people under a variety of "photometric inconsistencies" - changes in viewpoint direction, illumination and image distance. The measure of human performance will serve as a guide to the performance required by automated systems in order to compete with the accuracy levels of humans.
The two projects are expected to be completed in the next 18 months.
Press Releases & Announcements
This page last updated August 03, 2013