Report by Zhigeng Pan, Steve Cunningham and Judy Brown
One hundred and sixty delegates from twelve countries attended the VRAI 2002
(Virtual Reality and its Application in Industry) conference, April 9-12
in Hangzhou, Zhejiang Province, P.R. China.
Background of the conference
This conference, organized by the China Society of Image and Graphics,
provides an international forum for presentation and discussion of the
latest developments in fundamental research and practical applications
in virtual reality and its applications. The conference was co-organized
by the VR Committee of the China Society of Image and Graphics;
the State Key Lab of CAD&CG at Zhejiang University, China; and the
Hangzhou Institute of Electronics Engineering, China.
It was co-sponsored by EUROGRAPHICS, the VR Society in Japan (VRSJ),
Korean Association of VR (KAVR), IEEE-CS Beijing,
Natural Science Foundation of China (NSFC),
VR and Visualization Committee of China Computer Federation (CCF),
and Machine Perception and VR Committee of China Association of
Artificial Intelligence (CAAI).
VRAI 2002 is the third international conference on VR held in China.
The first one (International Workshop on VR and Visualization in Scientific Computing)
was held in 1995 in Hangzhou, China. Selected papers from this workshop were
published as a Special Issue of Computers & Graphics, Vol. 20, No. 2, 1996
(Guest Editors: Dr. Martin Göbel and Prof. Jiaoying Shi).
The second workshop (International Workshop on VR and Application) was held in 1999.
The plenary talks from this workshop were published by The International
Journal of Virtual Reality, Vol. 4, No. 3, 2000 (CD-ROM edition,
Content Editor: Prof. Zhigeng Pan).
In 2002, the conference focused on the theme
"VR and its Industrial Application,"
and selected papers will be published in a set of international journals,
such as Computers & Graphics
(Guest Editors: Prof. Jiaoying Shi, Prof. Zhigeng Pan),
and International Journal of Image and Graphics.
Hangzhou is extremely rich in natural and cultural
resources. There are two State-level scenic areas within the
city area: West Lake scenery area and Fuchun river-Xin'an
river - A-Thousand-Islet Lake scenery area. There is one
State-level nature preservation area: Tianmu natural
preservation area, two State-level forest parks:
A-Thousand-Islet forest park and Daqishan forest park. West
Lake, which is described as "Oriental Pearl", is located on
the east of the city proper, with the city on one side and
hills on the other three sides. Hangzhou, with its
surrounding hills and the famous West Lake, ranks as one of
the ten most scenic places in China. The conference
organizers arranged a half-day tour of West Lake for
overseas attendees and guests.
The conference opened with eight invited talks and closed
with four invited talks. Between these plenary sessions,
there were three parallel tracks that included papers on VR
algorithms, VR systems, VR Applications, User interaction in
VR, and distributed VR. Conference program co-chair Jiaoying
Shi chaired the opening session, beginning with brief
greetings from the Presidents of Zhejiang University and
Hangzhou Institute of Electronic Engineering, the conference
hosts. Martin Göbel, the other program co-chair, also
welcomed attendees and said a few words about the
conference. Figure 1 shows Martin Göbel, Program Co-Chair,
speaking during the opening session. Others seated at the
head table are, from left to right, Judy Brown, ACM SIGGRAPH
President, Yunhe Pan, President of Zhejiang University, Ming
Ye, Hangzhou Institute of Electronics Engineering, and
Jiaoying Shi, Zhejiang University and Program Co-Chair.
There were 165 paper submissions, with 66 accepted as
regular papers giving an acceptance rate of 40%. In
addition, there were twelve invited plenary speakers, thirty
short papers, and five tutorials.
At the end of the opening session, there was a formal
signing of the Cooperation Agreement between ACM SIGGRAPH
and China Society of Image and Graphics (CSIG.) Figure 2
shows Judy Brown, President of ACM SIGGRAPH, Yunhe Pan,
Chair of CSIG, and Jiaoying Shi, Vice Chair of CSIG signing
the agreement. Alan Chalmers, ACM SIGGRAPH Vice President,
was unable to attend and had signed it in advance.
Plenary speakers - first day
The first plenary speaker was Katsushi Ikeuchi,
University of Tokyo, speaking on Modeling from Reality.
There are many kinds of information that need to be captured
from reality in order to simulate it. This includes shape
information, photometric information (surface, texture, and
specularity), and environmental information. These must then
be merged into a consistent model for VR use, whether that
use is geometric or image-based modeling. Techniques were
presented for separating color information into specular and
lambertian color, as well as polarization techniques for
separating the specular and lambertian components of an
image. This is being applied to a number of cultural
heritage projects, including several large Buddhas in Japan,
Thailand, and China, as well as their historical
Judith R. Brown, ACM SIGGRAPH, spoke on Advanced Simulation and
Tele-Immersive Collaborations. As computing and network
technologies have advanced dramatically, the effectiveness
of simulations has increased, enabling innovations in
science, engineering, cultural heritage, distance education,
media communications, medicine, art, and architecture. Brown
focussed on applications, beginning with previous work done
by the Advanced Research Computing Services (ARCS) group she
had managed at The University of Iowa. She then presented a
variety of international tele-immersive collaborations from
the International Grid (iGRID) special event at the INET2000
examples were furnished courtesy of Maxine Brown, Electronic
Visualization Laboratory, University of Illinois, Chicago.
The talk featured the capabilities and applications of the
new U. S. National Advanced Driving Simulator, located at
The University of Iowa. (See
Harry Shum, Microsoft Research Asia, spoke on IBR: The Intersection of
Computer Vision and Computer Graphics. The presentation was
based on the author's SIGGRAPH 1999 paper on concentric
mosaics. The goal of this work is to reduce the dimension of
the image-based rendering problem, and concentric mosaics
can reduce this to four dimensions instead of the much
larger number of dimensions in an unstructured problem.
Martin Göbel, GMD, discussed Trends in VR Display
Technology for Industrial Applications. The holobench, or
responsive workbench, and several other examples of display
technology were described, including the iCONE, a
cylindrical system with sloping sides. Display technologies
included examples with olfactory display or sonification
effects. The results of olfactory display seem
disappointing, but they may have important use for special
applications such as hazardous material training. There is
little industry support for sonification, but it seems
important for telling stories or doing scientific work.
Different technologies have different values for
different kinds of work. Workbenches are seen to be better
than caves for science work, and two workbenches are a good
platform for collaborative work. However, caves seem to be
better for storytelling and education because they are more
immersive and provide more of an emotional experience. A
summary of projection systems with their fit for different
tasks was given.
Mark Green, City University of Hong Kong, presented
VR Authoring Tools
for Non-Programmers: A Case Study. There are great tools for
creative people to work in traditional media, but VR has
very few authoring tools for creative people. Green also
pointed out that, while these tools are very powerful, they
limit the creative capabilities of their users. For example,
students around the world are now learning packages like
Maya, and there are some aspects of traditional Asian design
that are being lost because these packages do not support
Effective VR tools need to support many features,
including object behavior; nonstandard devices; interaction
with a user, between users, and between objects; visual
richness; support for narrative and for scenes and
transitions; and multiple media. Green described current
work at the City University of Hong Kong on a scripting
system in which authors specify behaviors and programmers
define actions, and then any new action is available to all
the objects in the system.
Yi Tan, Princeton Video Image Corp., spoke about Virtual Imaging
and its Commercial Application in TV Broadcasting. Some
techniques were described that are now used to insert images
into television broadcasts (usually sports broadcasts) in
real time. These techniques involve feature extraction,
camera instrumentation, perspective correction, tracking,
and other processes. Similar processes could be used to
support augmented reality systems.
Kunwoo Lee, Seoul National University, presented Multi-Resolution
Modeling Operators for Components and Assemblies. In
developing a practical environment for collaborative design,
it is desirable to use level-of-detail techniques to adjust
representations of components or models. Two operators were
presented to simplify boundary representation models:
pruning the feature tree, to reduce the number of features
presented in the model, and wrapping an object, or replacing
the object by covering its concave regions.
Marcelo Zuffo, University of São Paolo, presented Cluster Computing for
Multiprojection Immersive Spaces. He described the current
state-of-the-art of commodity cluster computing for
multiprojection immersive spaces. Multiprojection immersive
spaces, such as CAVEs, have traditionally been built using
high-performance graphics systems. However, clusters of
commodity personal computers can provide the graphics
performance needed for these applications, and experiences
and examples were presented to show how this is done.
Advantages of such commodity clusters include low cost,
scalability, and flexibility. Figure 3 shows Marcelo Zuffo
outside the conference hotel.
Parallel paper sessions
On the second day, parallel paper sessions were held,
three parallel paper sessions in the morning and two in the
afternoon. There were a number of papers throughout the
parallel tracks that dealt with VR applications: "Virtual
Reality Enhanced Training for Minimally -Invasive Vascular
Intervention" by Yiyu Cai et al. from Nanyang Technological
University (NTU), Singapore was an example of applying
virtual reality and visualization techniques to medical
treatments. This gives an alternative to using a cadaver,
animals, or patients. Using haptic devices, the virtual
vascular patient can be modeled for any vascular disease and
can include intervention devices (catheters, guidewire,
balloons), motion tracking, and pressure tracking.
"EasyBowling: A Virtual Bowling System for Small Space" by
Weiwei Xu et al from Zhejiang University discussed the
implementation techniques for the EasyBowling system, which
has a real, physical bowling alley about six feet long and
uses a real bowling ball, the speed and direction of which
are tracked by a camera on the computer above the bowling
lane. The computer then determines which of the virtual pins
have been knocked down. Future work includes 3D realistic
audio rendering and internet-based game play. "The Design
and Implementation of Kyongju VR Theatre" by Hyoung-Gon Kim
et al. (Korea Institute of Science and Technology, KOREA)
provided another example of applying VR to entertainment.
In the distributed VR track, Lifeng Sun, Tsinghua
University, Beijing, China, discussed "Building
Collaborative Workshop of Metasynthetic Engineering Based on
Virtual Conferencing Space." Sun pointed out limitations of
desktop conferencing systems, defining the conference space
as the environment shared by all participants, and a space
model of awareness and interaction. Asame Alhalabi, Japan
Advanced Technology of Science and Technology, spoke on
"Network Latency Issue in a Cooperative Shared Haptic
Virtual Environment." He differentiated between a
collaborative haptic virtual environment, where sharing
users take turns manipulating and feeling a virtual object,
and a cooperative haptic virtual environment, where
interacting users simultaneously manipulate and act on the
same object at the same time. This kind of bi-directional
interaction is beneficial in medical tele-training,
telerobotics, tele-rehabilitation, and entertainment.
A reception, a banquet, along with lunch and coffee
breaks, gave ample opportunity for the attendees to meet
each other for informal discussions. Figure 4 is from the
banquet and shows, from left to right, Danniel
Thalmann, Haibin Wang, Sehwan Kim, HungJon Seo, ChangHoon
Park, and Seonhyung Shin.
Visit to Zhejiang University
Some of the participants visited the State Key Lab for
Computer Aided Design and Computer Graphics at Zhejiang
University. Students presented their research on morphing,
free form deformation, and non-photorealistic rendering.
This included hand-drawn animations and image-based
synthesis of Chinese landscape painting. Figure 5 shows
Qunsheng Peng and Jiaoying Shi at the State Key Lab.
Plenary speakers - third day
The third day included four plenary talks, an afternoon
sight seeing tour of west Lake, and the conference banquet.
Nadia Magnenat-Thalmann, University of Geneva, discussed
Physics Based Simulation
of Hair and Clothes. Modeling of the human body can be
time-consuming, but an approach was presented for "cloning"
an individual based on three photographs and a small set of
measurements. This creates bodies to H-ANIM standards from a
small set of general template bodies.
A number of problems in simulating hair and clothing are
seen to be approachable as physics problems. These include
interaction of hair with itself and self-collision for
clothing. Cloth and garment simulations may be done with
mechanical parametric approaches, and defining the
parameters of cloth can provide clothing that moves
Ruth Aylett, University of Salford, presented Intelligent Virtual
Environments. These are virtual environments that
incorporate areas of artificial intelligence in their
creation and/or operation. Adding this intelligence adds
context to help users understand what they are seeing so
they can interact in natural language, adds higher-level
representations of objects to add meaning to the scene,
makes the environment more dynamic and inhabited, and
provides higher-level interfaces with the users. Among the
other capabilities added by intelligence is the ability to
add agents with emotions and individual narrative to an
environment. However, artificial intelligence and computer
graphics have very different viewpoints, and there are a
number of research questions about the role of intelligence
in virtual environments. Application area driving this
technology include the computer games industry, e-commerce
(desire for talking heads type presenters), virtual
environment researchers seeking dynamics, and artificial
intelligence researchers seeking good testbeds. An agent
must have emotion to be believable because humans have
Soonhung Han, KAIST, Korea, discussed the Application of VR for Driving
Simulation: Virtual Bicycle Racing Simulator. In the VR
Institute at KAIST, researchers are developing simulators
for bicycle racing. These simulators are small, inexpensive,
with relatively simple mechanisms - and they are
sports-oriented and fun to use. One bicycle is mounted on a
6 degrees of freedom platform, but a newer example is
mounted on a 4 degrees of freedom platform (with yaw, pitch,
roll, and heave controls) that gives a much better
approximation to the measured behavior of a real bicycle.
Each simulator integrates four PCs to control the visual,
sound, motion, and dynamics aspects. Han suggested that
simulation sickness cannot be eliminated, but it can be
reduced by controlling parameters.
Dirk Reiners, FhG-IGD, Germany, closed the sessions with a presentation
on OpenSG: A Modern Open Source Scenegraph for Virtual
Reality Applications. Scene graphs are graphics systems
designed to permit much higher-level hierarchical design
than is supported by APIs such as OpenGL. OpenSG is a scene
graph that is designed to support design and development for
VR applications. It includes a number of features such as
multithreading, multiple pipe support, extensibility,
cross-platform operation, and handling free-form surfaces.
It is intended to fit into the same niche as Java3D and Open
Performer but to be open source and to have features that
fit better into the VR development world. It is intended to
be rolled out at the SIGGRAPH 2002 conference and to be
released in late 2002.
The conference organizers also organized a one and a half
day advanced seminar after the paper presentation sessions.
In the afternoon of April 11, the seminar was for domestic
participants and was given in Chinese. The courses on April
12 were given in English. The seminar included five courses:
"Recovering Geometry from Images" (Prof. Hongbin Zha, Peking
University, China), "Audio Rendering in Virtual Environment"
(Prof. Aiguo Song, South-East University, China), "Animation
Technique and its Application" (Prof. Nadia
Magnenat-Thalmann, University of Geneva, Switzerland; Prof.
Daniel Thalmann, EPFL, Switzerland), "Human animation"
(Prof. Hyeong-Seok Ko, Seoul National University, Korea),
and "Virtual Reality: Immersion, Decisions, Empathy"(Prof.
David Zeltzer, Fraunhofer CRCG, USA). The seminar was very
In addition, the ACM SIGGRAPH video "The Story of
Computer Graphics" was shown to the audience at the end of
A tour of the Yangtze River and the Three Gorges was
taken by some of the conference participants, immediately
following the conference. Much of this beautiful scenery
along the Yangtze will disappear after the largest dam in
the world is built across the Yangtze River. Completion date
for the dam is 2009.
To further internationalize this conference, the sponsor
is planning to hold the VRAI conference outside China.
Singapore may be a good choice, since attendees from Nanyang
Technology University are very interested in hosting it. For
additional information, see