Covise Collaborative Visualization and Simulation Environment is a framework for virtual reality use

1
Visualizing Image Data Sets with Covise
David Jackson and Ava Pierce Jacobs School of
Engineering, UC San Diego undergraduates
Introduction Covise (Collaborative Visualization
and Simulation Environment) is a framework for
virtual reality used and being refined in the
Visualization Lab at Calit2, UC San Diego. As a
participant in the NSF-sponsored Pacific Rim
Undergraduate Experiences (PRIME), we were able
to add functionality to Covise in the Cybermedia
Center at Osaka University, with the help of my
mentor Jurgen Schulze and through collaboration
with fellow PRIME student Ava Pierce (Fig
1). Scientists would like to use Covise to
intuitively manage and visualize large sets of
information, including very large and detailed
images and 3D models. They would like tools to
better express and analyze these objects.
Previous versions of Covise include a plugin that
can load images selected from a predefined set of
images, but no way to load any image on the
filesystem easily. Without a browser, users need
to manually code files they would like to load
into a configuration file, which is not useful
for a scientist that would like to manage large
sets of images. The goal of this project was to
create a thumbnail browser plugin that could be
incorporated with the image plugin, and Covise as
a whole.

Thumbnail Browser Interface Initial parameters
are stored in a global configuration XML file.
These can be edited to customize the browser for
specific displays. Single master node computers,
tiled wall displays, and immersive virtual
reality CAVE environments all require different
placements, orientations and aesthetic changes to
make the browser fit the interface. An auto-fit
option lets the plugin decide what the most
optimal location should be, but may not provide
the best solution in more complicated
environments. The thumbnail browser can be
accessed through Covises built-in menus. A
submenu allows the user to open the browser, and
toggle other functionalities (Fig 3). Once
opened, the browser dynamically loads converted
thumbnails of images in the current directory.
Pthreads are used to load the content in
parallel, allowing the user to manipulate the
browser and perform other tasks while it is
loading. The browsers geometry and dependent
positional matrices can be expressed through a
tree of Open Scene Graph nodes (Fig 4).
The browser itself can
initially be dragged horizontally, but can be
altered to move in two or three dimensions,
depending on the environment. Scrolling arrows,
when hovered over, move the browser horizontally
at a gradient, speed. These two forms of
interaction allow the browser to load the
complete directory of images at once and
onlydisplay the part that fit on screen, while
providing quick and adjustable access to them
all. Alternatively, the user can unlock the
browser and manipulate it in 3D world space.
Specific to loading images, additional support
has been added. When over a thumbnail, a tool
tip will display the name, size and dimensions of
the image. A histogram can also be generated,
displayed, and manipulated. When a thumbnail is
clicked, a message is passed to the image plugin,
which loads the corresponding osga image in an
intuitive way (Fig 5).
Figure 5. Work flow diagram of thumbnail browser
Future Work Researchers currently expanding
Covise have expressed desire to use this
thumbnail browser as it is. However, the
long-term goal of this plugin is to create a
universal, fully-functional browser for all
Covise applications, not just the image plugin.
A future version will be able to navigate up and
down through directory trees instead of just
pointing to one directory at the start. It will
also be able to recognize other types of files,
such as 3D protein models or videos, and pass
messages to respective plugins to load these
files. A singular navigation and visualization
system for all Covise files will allow
researchers and scientists alike intuitive access
to information in the system.
Figure 1. Tiled wall display running Covise at
the Cybermedia Center, Osaka University in Japan.
The thumbnail browser is loaded.
Figure 3. Standard layout for a master node
computer
Overview of Covise Covise can run on a variety
of interactive displays and environments.  It can
be used on a single computer with a mouse for
testing, but this does not take advantage of the
full immersive capabilities.  Covise is ideally
run on a tiled wall display or a CAVE with
additional devices such as 3D pointers, head
tracking, and stereo glasses which create a
completely immersive 3D experience. Covise uses
Open Scene Graph (OSG) for its 3D rendering,
which is an object oriented framework on top of
OpenGL. OSG also provides a node abstraction of
objects in the world. Images loaded by the image
plugin are in osga format. Different resolution
levels are stored to provide the best resolution
on screen, similar to zooming in on satellite
maps.
Conclusion This project provides a simple,
versatile way to manage and visualize file
systems through Covise. As Covise expands its
plugin library, and as scientists desire to
visualize larger and larger sets of data, the
browser implemented can be applied to more and
more projects.

• Acknowledgments
• Project in collaboration with Ava Pierce, PRIME
  2007 participant
• Research mentors
• Shinji Shimojo and Susumu Date, CyberMedia
  Center, Osaka University, Japan
• Jürgen Schulze, UC San Diego, California
  Institute for Telecommunications and Information
  Technology (Calit2)
• Prime advisors
• - Gabriele Wienhausen, Peter Arzberger, Linda
  Feldman and Teri Simas

(a)
(b)
(c)
Figure 2. Covise running in the StarCAVE at
Calit2, UCSD. Left Thumbnail Browser plugin.
Right Exploration through 3D model of Calit2.
Interaction is enhanced with head tracking (a),
stereo glasses (b), and a 3D pointer (c)
Figure 4. Scene Graph hierarchy of nodes for
thumbnail browser
For more information, please email David Jackson
at d2jackso_at_ucsd.edu. An electronic copy can be
found at prime.ucsd.edu or ieee.ucsd.edu