2 1541-1672/13/$31.00 © 2013 IEEE Ieee INteLLIGeNt SYStemS
Published by the IEEE Computer Society
CYBER-PHYSICAL-SOCIAL SYSTEMS
Editor: Liuqing Yang, Colorado State University, liuqingyang.ieee@gmail.com
IEEE 1857: Boosting Video
Applications in CPSS
Tiejun Huang, Yonghong Tian, and Wen Gao, Peking University
fl ows among the three spaces, video is de nitely
the one that takes the majority of traf c.
2
Yet,
video hasn’t received enough attention from the
CPSS community. Nowadays, millions of surveil-
lance cameras are being linked together via the
broadband network, consequently leading to a
“visual perception network” that’s instituted vir-
tually anywhere in the world. Such a world-scale
perception network offers moment-by-moment
pictures of both the physical world and human be-
havior over extended periods of time, providing a
panoramic digital mapping of the running state of
the world.
Obviously, there exists an enormous—and
growing—gap between the amount of video data
continuously collected by cameras and our ability
to ef ciently transmit, store, and intelligently ana-
lyze and digest this visual information. The IEEE
1857 standard released in June 2013, as the rst
standard that supports highly ef cient surveillance
video coding and objects-of-interest representa-
tion in the coding bitstream, can be used to nar-
row this gap, and consequently boost the various
video applications in CPSS.
Bigger and Bigger Video Data in CPSS
From the silver-surfaced copper Daguerreotype
plates introduced in 1839 to the electronic video
camera tube invented in the 1920s, humans began
a history of capturing and permanently recording
the physical world. In the 1960s, closed-circuit
televisions (CCTVs) and video recorders (VCRs)
were rst installed to monitor buildings, railways,
and other public infrastructures. Years later, video
surveillance systems appeared in banks and stores.
By the 1990s, home security systems allowed
users to remotely control a camera through a Web
interface.
Recently, as the hosting cities of the 2008 and
2012 Olympic games, Beijing and London show-
cased the utility of large-scale video applications by
deploying about 1 million surveillance cameras in
public area. Intelligent transportation systems (ITS)
is another typical CPSS eld in which video cam-
eras are employed to automatically monitor traf-
c fl ow, recognize vehicle license plates, detect vi-
olation behavior, and even identify people in the
crowd. Today, we notice surveillance cameras in
the elevator, on the ATM machine, along the sides
of streets, in of ce buildings, and almost anywhere
you can reach. A recent report from International
Data Corporation (IDC) shows that half of the
global Big Data in 2012 are surveillance video, and
the percentage will increase to 65 percent in 2015.
3
Thus, in terms of fusing computational, physical,
cognitive, and social domains, video is the most
important—and also possibly the dominant—data
that must be addressed in the CPSS study.
Challenges
For surveillance video, analysis can be done nor-
mally for purposes such as crime investigations,
military intelligence, or consumer traf c pattern
discovery. But from a CPSS viewpoint, surveil-
lance video contains a great deal of information
about society’s operations day and night. More-
over, what began with manually monitored video
displays has evolved into a variety of systems
with automated processes to scan multiple video
streams, detect events or objects of interest, and
act on them. In the future, more intelligence will
be embedded in the cameras and the surveillance
cloud to automatically analyze what’s happening
in “video big data.”
Technologically, there are many challenges in-
volved in using CPSS to process video big data.
Some of the foremost challenges involve how to
T
ypically, a cyber-physical-social system (CPSS)
features tight integration and coordination
among computational (or cyber), physical, and hu-
man (social) elements.
1
In all types of information
IS-28-05-CPSS.indd 2 11/12/13 12:40 PM