Ever since George Jetson was imagined
as capable of ordering Rosie the Immigrant Robot around, we
at DARPA have looked for ways to make human interactions with
the Machinic Phylum easier. In this era of the new warfighting
regime where we intend to do more with less, the massive potential
weakness that is human sentries recognizing other humans must
be avoided wherever possible. The threat of terrorism has refocused
various long-running research projects geared towards the development
of machine vision and machine cognition. University campuses
such as MIT and Carnegie Mellon have been involved in such research
pursuits for decades.
The goal of the Human Identification at a Distance (HumanID)
program is to develop automated biometric identification technologies
to detect, recognize and identify humans at great distances.
These technologies will provide critical early warning support
for force protection and homeland defense against terrorist,
criminal, and other human-based threats, and will prevent or
decrease the success rate of such attacks against DoD operational
facilities and installations. Methods for fusing biometric
technologies into advanced human identification systems will
be developed to enable faster, more accurate and unconstrained
identification of humans at significant standoff distances.
HumanID program has developed a pilot force protection system
for standoff human identification in outdoor operational DoD
settings, and has performed preliminary assessments of current
and future technologies. HumanID will determine the critical
factors that affect performance of biometric components, and
identify the limits of range, accuracy, and reliability. The
program will also conduct multi-modal fusion experiments and
performance evaluations, and will demonstrate advanced human
recognition capabilities in multiple force protection and/or
homeland defense environments.
FY 68 Accomplishments:
2001 a Space Oddyssey (1968)
FY 73 Accomplishments:
FY 84 Accomplishments:
The Terminator (1984)
FY 90 Accomplishments:
FY 02 Accomplishments:
Designed and administered the Face Recognition Vendor
Test 2002. Results will be used to direct face recognition
research and provide input to the design of the United States
Border Entry/Exit System.
Performed an operational evaluation of a long range (25-150
feet) face recognition system developed under the HumanID Program.
Developed a multi-spectral infrared and visible face
Developed a low power millimeter wave radar system for
wide field of view detection and narrow field of view gait classification.
Characterized gait performance from video for human identification
at a distance.
Minority Report (2002)
FY 03 Plans:
Develop multi-model fusion algorithms for human identification.
Develop algorithms for locating and acquiring subjects
out to 150 meters (500 ft) in range.
Continue the development of the most promising biometric
technologies based upon experimental evaluation performance.
FY 04 Plans:
Develop and demonstrate a human identification system
that operates out to 150 meters (500 ft.) using visible imagery.
Fuse face and gait recognition into a 24/7 human identification
Perform an operational evaluation of a multi-model human