When delegates of the 16th National Congress of the Communist Party
of China were lining up to enter Beijing's Great Hall of the People
on November 8, an automatic identity recognition system was used to
capture delegates facial information on camera before allowing them
to enter the hall without any delay. This facial feature detection
and recognition technology, developed independently by scientists
from the Computer Institute of the Chinese Academy of Sciences (
CAS),
passed formal expert appraisals in Beijing on November 26.
Research into the facial feature detection and recognition
technology started in 1994 as part of China's vital March 1986
Program. A person in charge of the Computer Institute of CAS says
the successful development of this core technology is a significant
breakthrough for the Program.
An
appraisal committee comprised of more than ten scientists including
academics Wang Chengwei, from the Chinese Academy of
Engineering, and Tan Tieniu, head of the Automatic Institute of
CAS, concluded after serious research and discussion that the
facial feature detection and recognition technology has made
breakthroughs in face detection, facial features extraction, face
recognition and verification. It is the leading technology in China
and has reached advanced international standards.
Gao Wen, chief scientist from the digital research group of the
Computer Institute of CAS and the project's leading researcher,
gave a demonstration of their works. The automatic identity
recognition system is able to capture the facial information of an
unidentified person via a camera and analyze their facial features,
including skin color, veins, and shape, before comparing them with
parameters saved in the system's database. The identity of the
person will be recognized immediately.
"It takes only one-tenth to one-twentieth of a second to detect a
person's face and one minute to provide recognition of a person's
identity on an ordinary personal computer. It will be even faster
on better equipped computers," said Gao Wen, "Human physical
feature recognition technology including the recognition of facial
features, fingerprints, palms, irises, voices, signatures and
retinas, are all computer security technologies that have been
developed in recent years. As a noninvasive form of recognition
technology, the facial feature recognition system is an important
scientific technique that can be applied in anti-terrorism and
security."
"Compared with the best current system of the same category in the
US, our product has more advantages in certain situations such as
being able to perform functions against complicated background,
under gloomy light and shaky condition," said Gao Wen.
It
is reported that the facial feature recognition technology has
played a special and important role in many areas including police
surveillance, jail monitoring, judicial identification, civil
aviation security, port monitoring and providing verification of
various identification papers and bank card holders.
(China.org.cn by Wang Qian, December 3, 2002)