Transcript VIS 2004 - CERN Indico

IEEE Visualization 2004
Sponsored by IEEE Computer Society Technical Committee on Visualization and Computer Graphics in
Cooperation with ACM/SIGGRAPH.
Visualization 2004 is the premier forum for visualization advances in science and engineering for academia,
government, and industry. This event brought together researchers and practitioners with a shared interest
in techniques, applications, and technology.
Visualization 2004 was held in Austin, Texas, October 10-15, 2004.
Web: http://vis.computer.org/vis2004
Program
The program included:




Tutorials
InfoVis 2004: IEEE Symposium on Information Visualization
VolVis 2004: IEEE/SIGGRAPH Symposium on Volume Visualization and Graphics
Conference
Tutorials
>>>
>>>
>>>
>>>
>>>
>>>
>>>
Tutorial
Tutorial
Tutorial
Tutorial
Tutorial
Tutorial
Tutorial
1:
2:
3:
4:
5:
6:
7:
Information Visualization and Visual Discovery
Procedural Encoding of Scattered Data, Theory and Applications
GPGPU: General Purpose Computation on Graphics Processors
Advanced Virtual Medicine: Techniques and Applications for Medicine-oriented Visualization
An Introduction to Visualization Using VTK
Interactive Texture-Based Flow Visualization
Feature Oriented Methods in Flow Visualization
Tutorial 1
Information Visualization and Discovery
Georges Grinstein ( University of Massachusetts Lowell),
Daniel Keim (University of Konstanz),
Matthew Ward (Worcester Polytechnic Institute).
This tutorial provided the necessary background to understand the issues in the development and usage of visualization
integrated with data mining and knowledge discovery systems. It provided a brief history of data visualization and data
mining and examined both sample commercial and academic knowledge discovery systems that integrate visualization and
data mining. Many slides (350), videotapes and demonstrations were provided.
Goals of Tutorial:
•
Look at history of Visualization and Visual Data Mining
• Understand the issues in interactive data visualization
• Examine numerous visualization techniques and systems
• Look at a specific application example area (bioinformatics, …)
• Explore the future of visualization
Tutorial 1
A Definition of Visualization:



It is the Visual Interface to the data and mining tools
It is a method of interacting with data and algorithms

supports the user through all the knowledge discovery steps

uses selections, queries, probes, and view transformations
It is completely separable from the analysis methods

Data can be analyzed using many different algorithms

Each result can be viewed in a different visualization

Each visualization thus provides a different view of the results
Tutorial 1
A Definition of Visual Data Mining:
Data Mining is the process of searching and analyzing databases to find implicit but
potentially useful information
What are the Key Data Factors ?
•
Very large number of parameters: more than 100
• Very large data sets: more than 10**7
• Multiple data types: discrete and continuous
• Noisy data: often not uniform
• Missing values: could be important
• Lots of different tasks
These numbers depend on the domain and increase with time (in a couple of years
they will be shift)
What is High Dimensional?
Tutorial 1
Goals of Visualization Techniques

Presentation




Confirmatory Analysis





Facts to be presented are fixed a priori
Need to select appropriate presentation techniques and parameters
Result is a high-quality visualization to present the known facts
Have some close to specific hypotheses about the data
Goal-oriented examination of these hypotheses
Result is a visualization of data that confirms, rejects or provides
more information on the hypotheses
Exploratory Analysis





Have no hypotheses or very broad ones about the data
Need to explore interactively, usually as undirected searches for
structures, trends, ...
Result is a visualization of data that hopefully leads to hypotheses
about the data
Tutorial 1
Visualization Techniques :
Pure
4 2D and 3D Scatter plots
4 Matrix of Scatter plots
4 Statistical Charts
4 Line and Multi-line Graphs
4 Parallel Coordinates
4 Circle Segment
4 Polar Charts
4 Survey Plots
4 Heatmaps
4 Height Maps
4 Iconographic Displays
4 RadViz and PolyViz
Integrated with Analysis
4 Projection Pursuit
4 Dimensional Stacking
4 Sammon Plots
4 Multi-Dimensional Scaling
4 PCA and Principal Curves
4 Self Organizing Maps
Interactions
4 Selection
4 Probing, Querying
4 Grand Tours
4 Non-linear Zooms
Modern Visualization Techniques
Tutorial 1
Classic Visualization Techniques
Standard 2D/3D Displays
Ultra Modern Visualization Techniques
Tutorial 1
Visualization can be misleading : colors, maker shapes
shading should be used carefully. The type of representation
too. For instance histograms are better than pie chart
because our eyes are very bad at angles differentiation.
Florida 2000 Election Recount:
Count and total black dots for Al Gore and white
dots for George Bush. Recount to confirm.
Tutorial 2
Procedural Encoding of Scattered Data, Theory and Applications
Greg Nielson (Arizona State University),
Hans Hagen (University of Kaiserslautern),
David Ebert (Purdue University),
Kelly Gaither (The University of Texas at Austin).
“Procedural encoding of scattered data sets is an active area of research with great potential for reconstructing surface
information and compactly representing large data. The reduced storage requirements allow greater flexibility in the
methods for manipulating and analyzing this data interactively. This course, covered both the mathematical foundations
behind existing encoding techniques, surface reconstruction methods, and volumetric representations. Additionally, it
presented methods for feature analysis in the functional domain and concluded with applications and benefits of functional
encoding in the scientific and engineering disciplines.”
What Attendees Gained:
The mathematical foundations for encoding scattered,
surface and volumetric data and concrete examples of
extracting features from and visualizing the data expressed in a
functional representation.
Tutorial 5
An Introduction to Visualization Using VTK
Lisa Avila (Kitware)
Berk Geveci (Kitware)
William Schroeder (Kitware)



This tutorial was designed to introduce attendees to the most common visualization techniques and provide
concrete examples of these techniques using The Visualization Toolkit. It was designed for attendees who have
some programming knowledge but are not experts in the field of visualization.
The focus was on geometric, scalar, and vector field visualization using techniques such as iso-surfacing, direct
volume rendering, streamlines, scalar coloring, glyphing, and image processing.
The Visualization ToolKit (VTK) was used a base to demonstrate the concepts.







VTK is an open source, freely available software system for 3D computer graphics, image processing, and visualization.
VTK consists of a C++ class library, and several interpreted interface layers including Tcl/Tk, Java, and Python.
VTK supports a wide variety of visualization algorithms and advanced modeling techniques.
VTK has been installed and tested on nearly every Unix-based platform and PCs (Windows 98/ME/NT/2000).
VTK is interfaced to OpenGL.
www: http://www.kitware.com/vtk/
Even if it was a “beginner” level, it was an interesting talk and a good overview of all basics graphics concept.
INFOVIS
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
Keynote Address: From Information Visualization to “Sense making”
Session 1: Evaluation
Session 2: Techniques and Tools
Session 3: Multivariate Data
Session 4: Design Studies
Session 5: Evaluation and Perception
Session 6: Models and Systems
Session 7: Graphs
Capstone Address: Visualization as a Medium for Capturing and Sharing Thoughts
Posters
INFOVIS: Keynote Address
From Information Visualization to Sense making: Connecting the Mind's Eye to the Mind's
Muscles
Stuart Card (Palo Alto Research Center PARC)
Human-Computer Interaction expert



Understand better how the mind and eyes is working.
Example application: Book reading (a book emulation on screen).
Info visualization is done (solved) next problem is sense making
“Looking back: On the tenth anniversary of the first Symposium on Information Visualization, much has been
accomplished. The computer's power has been exploited to give quick visual form to abstract data, to interact, and
to warp detail to follow the user's changing interest. Moreover, the design space of visualizations has been
systematized with reference models, taxonomies, and monographs relating visualization to perceptual and graphical
constraints.”
“Looking ahead: The era of pure information visualization is over. The path ahead depends on giving much more
attention to the purposes of visualization and its use. Leaving aside communication, the purpose of
information visualization is insight, or more particularly, a larger process that might be called sense making.”
INFOVIS: Techniques and Tools
Geotime Information Visualization
Thomas Kapler and William Wright (Oculus Info Inc)
(OCULUS tool) geographical and time information combined . Nice GUI and 3D plots. Demo presented the author
life. Show graphs on earth. Places according to time.
INFOVIS: Techniques and Tools
RecMap: Rectangular Maps approximation
Roland Heilmann (Bayer Technology), Daniel Keim, Christian Panse, and Mike Sips (Konstanz University)
Rectangular cartograms. Approximate a map to rectangles proportional to data . Set of polygons, and a value for
each of them. The result is a sets of rectangles proportional to the statistical value. (Ex: US 2000 elections map)
(b) RecMap for (MP1)
(a) Traditional map
Visualizations of the presidential race in 2000. The areas of Figures (b)
and (c) correspond to the numbers of electoral voters. The red and the
blue color depict which candidate has won each state. The candidate who
covers most of the area of all polygons (b) and (c) has won the election.
(c) RecMap for (MP2)
INFOVIS: Techniques and Tools
EZEL: a Visual Tool for Performance Assessment of Peer-to-Peer File-Sharing Networks
Lucian Voinea, Alex Telea, and Jarke van Wijk (Technische Universiteit Eindhoven)
EZEL is a visual tool for the performance assessment of peer-to-peer file-sharing networks using a new visualization
technique, faded cushioning, which allows visualizing the same data from different perspectives.
Visual tool for the assessment of
performance in P2P file-sharing
networks (EZEL snapshot)
INFOVIS: Techniques and Tools
History mechanism for visual data mining
Matthias Kreuseler (SD Industries GmbH), Thomas Nocke, and Heidrun Schumann, (University of Rostock)
A major challenge of current visualization and visual data mining (VDM) frameworks is to support users in the orientation
in complex visual mining scenarios. An important aspect to increase user support and user orientation is to use
a history mechanism that, first of all, provides un- and redoing functionality. This paper presented a new
approach to include such history functionality into a VDM framework.
Screen shot from the VDM System InfoVis (a demographic
data set of the countries of the world) with History tree
(left), DataTable display in a table lens mode (center), tree
visualization (MagicEyeView) of result of a hierarchical
cluster algorithm (top-left), parallel coordinate display with
the selected polyline of Botswana (center-bottom), and a
ShapeVis display (bottom-right) showing all the countries
from the captured countries from the MagicEyeView
INFOVIS: Multivariate Data
Steerable, Progressive Multidimensional Scaling
Matt Williams and Tamara Munzner (University of British Columbia)
Current implementations of Multidimensional Scaling (MDS), an approach that attempts to best represent data
point similarity in a low-dimensional representation, are not suited for many of today's large-scale datasets
(too slow). This paper proposed an extension to the spring model approach that allows the user to interactively
explore datasets that are far beyond the scale of previous implementations of MDS. (1st streerable MDS algorithm)
Left: 40,000 point real environmental
dataset laid out with the Morrison
algorithm, taken after a full layout
computation that takes 16 minutes.
Right: We show a partially placed version
of the same environmental dataset after
steering with MDSteer for roughly 2
minutes. Again, we see the same
large-scale structure.
INFOVIS: Multivariate Data
A Rank-by-Feature Framework for Unsupervised Multidimensional Data Exploration
Using Low Dimensional Projections
Jinwook Seo and Ben Shneiderman (Department of Computer Science & Human-Computer Interaction Laboratory, Institute for Advanced Computer Studies University of Maryland)
HCE hierarchical clustering Explorer.
Goal: find features in Multi Dim Data Set. All the possible scatter plot of the data set (ntuple) are visible at the same
time (small). The correlation coefficient is shown as a color plot.
INFOVIS: Multivariate Data
Value and Relation Display for Interactive Exploration of High Dimensional Datasets
Jing Yang, Anilkumar Patro, Shiping Huang, Nishant Mehta, Matthew O. Ward, and Elke A. Rundensteiner (Computer Science Department Worcester Polytechnic Institute)
A new multi-dimensional visualization technique named: Value and Relation (VaR) display. The VaR display
uses pixel-oriented techniques to utilize screen space efficiently. By graphically presenting each dimension of a
high dimensional dataset as a glyph in a 2D space, the VaR display conveys relationships such as correlation among
the dimensions through the positions of the glyphs. The positions of the glyphs are generated using Multidimensional Scaling (MDS) according to the pair wise relationships among the dimensions. MDS is a technique
that maps locations in high dimensional space to positions in a low dimensional space.
Figure a, each dimension of the SkyServer dataset (361
dimensions, 50,000 data items) is mapped to a dot and
positioned in the 2D space using MDS. Such a display is
called a star field display. In this example the
correlation among the dimensions is used to generate
the positions through MDS. Thus closely related
dimensions have positions adjacent to each other
in this display. It reveals the correlation among
the dimensions intuitively.
Figure b shows the VaR display of the SkyServer
dataset generated by replacing the dots in Figure 1 by
glyphs. The textures of the glyphs reveal data
patterns in the dimensions.
INFOVIS: Multivariate Data
Uncovering Clusters in Crowded Parallel Coordinates Visualizations
Almir Olivette Artero, Maria Cristina Ferreira de Oliveira (Department of Computer Science University of São Paulo),
Haim Levkowitz (Department of Computer Science University of Massachusetts)
Parallel Coordinates visualization of a data set with 7,500 five-attribute records. This data set has 2,920 of its records
distributed into five clusters, with Clusters 1 to 5 having 848, 728, 8, 608 and 728 records, respectively. The remaining
4,580 records were generated randomly, constituting noise. Nevertheless, it is impossible to observe the
clusters in the visualization.
The strategy consists in
computing frequency and
density information from
the data set, and using such
information in Parallel
Coordinates visualizations
to filter out the information
to be presented to the user,
thus reducing visual clutter and
allowing the analyst to observe
relevant patterns in the data.
INFOVIS: Multivariate Data
Clutter Reduction in Multi-Dimensional Data Visualization Using Dimension Reordering
Wei Peng, Matthew O. Ward, and Elke A. Rundensteiner (Computer Science Department, Worcester Polytechnic Institute)
In many multivariate visualization techniques, such as parallel coordinates, glyphs, scatter plot
matrices and pixel-oriented methods, dimensions are positioned in some one- or twodimensional arrangement on the screen. Given the 2-D nature of this medium, some ordering or
organization of the dimensions must be assumed. This organization can have a major
impact on the expressiveness of the visualization.
In order to automate the dimension reordering process for a display, three issues should be
considered:
1)
2)
3)
determining the way clutter manifests itself in the display,
designing a metric to measure visual clutter,
arranging the dimensions for the purpose of clutter reduction.
…/…
INFOVIS: Multivariate Data
INFOVIS: Multivariate Data
INFOVIS: Design Studies
Time-Varying Data Visualization using information Flocking boids
Andrew Vande Moere (Key Centre of Design Computing and Cognition, University of Sydney)
“This
research demonstrates how principles of self-organization and behavior simulation can be used to represent
dynamic data evolutions by extending the concept of information flocking, originally introduced by Proctor &
Winter, to time-varying datasets. […] Model the movements of so-called boids (or bird-objects) within a
flock.”
Few images:
INFOVIS: Design Studies
Artifact of the presence Era. Using information visualization to create and evocative souvenir
Fernanda B. Viégas, Ethan Perry, Ethan Howe, Judith Donath, (MIT Media Lab)
History of how a space (room) is used. Institute of contemporary art in NY. Inspiration from geology layer
(stones). Camera and microphone running continuously in the museum entrance. A wave every 5 minutes. The
shapes of the waves come from the sound and the color from the video. The bottom of the stack is
compressed like sedimentary stones.
INFOVIS: Capstone Address
Visualization as a Medium for Capturing and Sharing Thoughts
Steven Roth, (MAYA Viz, Ltd and Carnegie Mellon University)
One of the Founder of InfoVis (it was the 10th InfoVis anniversary).
Presented a tool to manage operations in Iraq. 2D, very interactive, many meta data, picking, moving object around
(background is a map). Task management combined with geographical information.
His messages were:
•
•
•
Visualization alone is now a well defined area
Need to evolve the visualization field by making more sense to visualization
Visualization useless without context application
VOLVIS
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
Keynote Address: Quo Vadis, VolVis?
Session 1: Evaluation
Session 2: Techniques and Tools
Session 3: Multivariate Data
Session 4: Design Studies
Session 5: Evaluation and Perception
Session 6: Models and Systems
Session 7: Graphs
Capstone Panel: What is the future of Volume Visualization & Rendering?
VOLVIS: Capstone Panel
1996 Retrospective Panel – What is the Future of Volume Visualization & Rendering
Moderator: Claudio Silva (University of Utah)
Panelists: Chris Johnson (University of Utah), Randall Frank (CEI International), Michael Meissner (Viatronix Inc), Michael Doggett (ATI)
The only VolVis I followed . Sort of summary more interest in InfoVis (room full).
No note provided (panel) – 15 minutes/panelist
Chris Johnson: Multi modal (Ex: a picture from MRI and scan) - Multi valued rendering (module render in several ways at the same time) – Multi
field visualization – Accuracy issues and uncertainty quantification - Time optimized acquisition & visualization (future of genetics is
imaging) – Perceptual issues – Theory of volume visualization.
Randall Franck : We’ve come a long way (since 96) performance increased a lot. Rectilinear grid, rendering quality, flexibility, great progress in
unstructured rendering. So why uneasy feeling ? Still using the same conceptual engine written in 94 (end users very happy with) –
Volume visualization tools must be usable by non graphics users – When to use volume rendering ? – Large data sets are particular – Is
information being conveyed here ? – Continuing challenge : time varying data, unstructured volume .
Michael Meissner: A long way, some acceptance 3 or 4 speakers are from industry – Algorithms: in the future image quality progress will be
less obvious -Technological improvements is huge but the volume of data increases too: the gap will remain – Pretty image doesn't mean
meaningful – Meaningful is defined by the application – Do we need the 3d visualization ?(medical application) – Interaction too
complicated - Volume visualization is a communication toolbox – Future topics : Large data sets – Non photorealistic volume rendering – Multi
modal visualization – 4D – Visualization of segmented volumes.
Michael Dogget: Graphics hardware centric – Move more rendering to the GPU – Virtual video memory (requirement for longhorn) – More flexible
memory management - More application driven – User involvement – Non photo realistic rendering GPUS (make volume visualization
really used, not just a fancy thing cf: medical field)
Audience: Missing big challenges (Something which is impossible to do in volume visualization now) - Are they too far away from Real Science ? –
Scale problem (community too small) ?
VIS
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
Keynote Address: The Human Visual System: How Is Its Design Related to the Physics of the Natural Environment?
Session 1: Volume Rendering
Session 2: Flow Visualization I
Session 3: Volume Rendering II
Session 4: Flow Visualization II
Session 5: View/Importance Driven Rendering
Session 6: Systems
Session 7: Flow Visualization III
Session 8: Meshes
Session 9: Molecules I
Session 10: Surfaces
Session 11: Lighting and Perception
Session 12: Flow Topology
Session 13: Large Data
Session 14: Medical Visualization I
Session 15: Finite Element Analysis
Session 16: Isosurfaces I
Session 17: Molecules II
Session 18: Isosurfaces II
Session 19: Medical Visualization II
Session 20: Real World
Capstone Address : Self-Illustrating Phenomena
Posters
VIS: Keynote Address
The Human Visual System: How is its design Related to the Physics of the Natural
Environment
Wilson S. Geisler (Center for Perceptual Systems and Department of Psychology University of Texas at Austin)
“The human visual system is the result of evolution by natural selection and hence its design must incorporate detailed
knowledge of the physical properties of the natural environment. This is an obvious statement, but the scientific
community has been slow to take it seriously. Only recently has there been an increased effort to directly measure
the statistical properties of natural scenes and compare them to the design and performance of the human visual
system.”
This presentation described some recent studies of the chromatic and geometrical properties of natural
materials and natural images, as well as some perceptual and physiological studies designed to test how
those physical properties are related to human perceptual mechanisms.
VIS: Systems
Visualization in Grid Computing Environments
Ken Brodlie (Leeds University)
This research, evolves the dataflow concept to allow parts of the visualization process to be executed remotely in a
secure and seamless manner.
Extended distributed Iris explorer using XML (sKML = generic visualization description) – Grid computing.
VIS: Systems
Visualizing Competitive Behaviors in Multi Virtual Environments
Nate Hoobler (Virginia University)
A system for enhancing observation of user interactions in competitive virtual environments (video games).
VIS: Systems
Scout: A Hardware-Accelerated System For Quantitatively Driven Visualization and Analysis
Patrick A.McCornick (Utah University)
Many visualization techniques rely on indirect mappings, such as transfer functions, to produce the final imagery. In many
situations, it is preferable and more powerful to express these mappings as mathematical expressions, or queries, that
can then be directly applied to the data. This paper presented a hardware-accelerated system that provides such
capabilities and exploits current graphics hardware for portions of the computational tasks that would
otherwise be executed on the CPU. The direct programming of the graphics processor using a concise data
parallel language, gives scientists the capability to efficiently explore and visualize data sets.
VIS: Surfaces
Very technical session, a bit outside our field but was interesting to see.
Rough interface reconstruction using level set methods
Yootai Kim (Ohio State UNI)
Ex: ice on aircraft wings
Surface reconstruction of noisy and defective data sets
Hui Xie (siemens corporate research)
Incomplete data sets – Algorithm to handle noisy and not
continuous surfaces (seed growing)
VIS: Surfaces
Optimal global conformal surface parameterization
Yalin Wang: Sunny, UCLA, Harward UNI
This paper provides an explicit method for finding optimal global conformal
parameterizations of arbitrary surfaces.
Local and global comparison of continuous functions
Vijay Natarajan UC Davis
The topic of this paper is a new comparison
measure for functions defined on a common
Riemannian manifold.
VIS: Large Data
Interactive Exploration of Large Remote Micro-CT Scans
Stephan Prohaska, Zuse Institute (Germany)
The article described the design of a visualization system allowing scientists to interactively explore large remote
data sets in an efficient and flexible way. The system is broadly applicable and currently used by medical
scientists conducting an osteoporosis research project. Human vertebral bodies are scanned using a high resolution
micro-CT scanner producing scans of roughly 8 GB size each. Due to the rich internal bone structure, scientists
need to interactively explore the full dataset at coarse levels, as well as visualize sub volumes of interest at the
highest resolution.
VIS: Large Data
Interactive terascale particle visualization
David Ellsworth Nasa
This paper describes the methods used to produce an interactive visualization of a 2 TB computational fluid
dynamics (CFD) data set using particle tracing (streak lines).
VIS: Large Data
Intuitive and interactive modification on large finite element models
Dirc Rose, Katrin Bidmon, Thomas Ertl (Visualization and Interactive Systems Group, University of Stuttgart)
Interactive algorithms for intuitive and fast editing of Finite Element models and appropriate visualization
techniques to support engineers in understanding these models.
This accelerate the development cycle: simulation
engineers are able to modify their Finite Element models
without going back to the CAD department.
VIS: Large Data
Visualization of Salt-Induced Stress Perturbations
Patricia Crossno, David H. Rogers, Rebecca M. Brannon (Sandia National Laboratories) David Coblentz (Los Alamos National Laboratories)
Abstract:
“An important challenge encountered during post-processing of finite element analyses is the visualizing of threedimensional fields of real-valued second-order tensors. Namely, as finite element meshes become more complex
and detailed, evaluation and presentation of the principal stresses becomes correspondingly problematic. In this
paper, we describe techniques used to visualize simulations of perturbed in-situ stress fields associated with
hypothetical salt bodies in the Gulf of Mexico. We present an adaptation of the Mohr diagram, a graphical paper
and pencil method used by the material mechanics community for estimating coordinate transformations for stress
tensors, as a new tensor glyph for dynamically exploring tensor variables within three dimensional finite element
models. This interactive glyph can be used as either a probe or a filter through brushing and linking.”
VIS: Panel 1
Can we determine the top unresolved problems of visualization
Organizer Theresa-Marie Rhyne TMR; Panelists: Bill Hibbart BH, Chris Johnson CJ; Chaomi Chen CC, Steve Eick SE.





BH: Top 5 main problems : #5 Big data - #4 Interaction - #3 Pretty pictures (labels, good density of
information, printing (PS PDF)) - #2 Abstraction - #1 Finding funding - Scientist (the users) are busy with
their science have no time to spend with visualization experts.
CJ: #1 Think about the science - Accuracy and uncertainty (making sense again) – Multi field imaging and
visualization – Time dependent visualization (interaction with ..) – Scalable visualization – Can we get beyond
linear methods ? –
CC: User attention may be drawn to false patterns unintentionally – Interdisciplinary – InfoVis / SciVis –
Identifying emerging trends .
SE: 1) Visual scalability - 2) Visualizing new data structure – 3) Application to new problems – Being
overwhelmed by information: bigger display, tile display, reduce amount of data ... - Factor affecting
scalability: human perception, (close to the maximum resolution human can see ?) .
Audience: “We (vis experts) do not have customers….” “software created are mainly used by visualization
researchers” “Significant needs because of the new data around (GPS, medical, etc …)” “Vis cannot be seen as an
isolated thing” “Feed back the Vis techniques in the real world (science)” “Do not have customers because real
data are huge and the techniques do not scale”
VIS: Panel 2
In the Eye of the Beholder: the role of perception in scientific visualization
David Ebert DE Purdue UNI; Kelly Gaither KG UT Austin; Bill Geisler BG UT Austin; David Laidlaw DL, Brown Uni




DE: Effective Information Transmission – so much data – abstract away unimportant details – Utilize illustration
principles and techniques – Perception research – Advanced illumination and shading
BG: Human visual system is able to infer shapes –> contrast effect can produce dramatic effect
DL: All visualization is perception – Roles of perception: using knowledge from perceptual psychology -> no
red text on blue background, visual working memory, contour detection etc … – Methodology form perceptual
psychology - Knowledge form art and design - Artists may help to do good visualization.
Audience: “increase the number of artist and designers in that environment as well as users” “Training might affect
how people see things” “Having interactive control on what you see as effect on the way you see it” – “Looking
natural or not ?”
VIS: Panel 3
The Future Visualization Platform
Chair Greg Johnson GJ: Austin UNI; Panelists: David Ebert BE, Purdue UNI; Charles Hansen CH, Utah UNI; David Kirk DK, nVIDIA Corp; Bill Mark BM, Austin UNI; Hanspeter Pfister HP,
Mitsubishi.






BE: Hardware direction: increase power (4 frame/sec with all the shading etc …) With these performances there is still
issues: What to show ? How to show it ? How to interact ? User potential uses: interactive simulations running on
the GPU. Use GPU as mush as possible.
CH: Interactivity is the key.
DK: GPU Visualization servers – GPUS extremely powerful . GPUS continue to accelerate above Moore's law
no comparison with CPUs
BM: Easy to use, reliable, high performance, permit to focus on task - Despite progress not yet there - GPU are very
parallel but very specialized CPU are very general but sequential but they are moving to a common point
–– In 6 years One processing ships ?
HP: Graphics hardware and rendering are becoming commodities - CPU enough for mass market: GHz replaced by
added value – Touch table: new desk ? - Future success depend on conveying information effectively.
Audience: “Nobody mentioned the data rate produce by science ?” “ Close collaboration between vis scientist and
domain scientist.” “Do not confuse data and information” “faster CPU and GPU but what about memory speed ?”
Vis: Isosurfaces I
Efficient Point-Based Isosurface Exploration Using the Span-Triangle
Bartosz von Rymon-Lipinski Nils Hanssen Thomas Jansen Lutz Ritter Erwin Keeve (Surgical Systems Laboratory, research center caesar)
A novel span-triangle data structure, based on the span-space representation
for isosurfaces. It stores all necessary cell information for dynamic
manipulation of the isovalue in an efficient way.
Volume Refinement Fairing Isosurfaces
Martin Bertram (TU Kaiserslautern)
An interpolating refinement method for two- and
three-dimensional scalar fields defined on hexahedral grids.
Interactive Point-Based Isosurface Extraction
Yarden Livnat Xavier Tricoche (Scientific Computing and Imaging Institute, University of Utah)
A novel point-based approach to view dependent iso-surface
extraction.
Vis: Real World
Interactive Design of Multi-Perspective Images for Visualizing
Urban Landscapes
Augusto Román, Gaurav Garg, Marc Levoy (Computer Graphics Laboratory, Stanford University, CA)
This paper describes an interactive system for creating multi-perspective images composed
of serially blended cross-slits images.
Real-Time Motion Estimation and Visualization on Graphics Cards
Robert Strzodka (caesar research center), Christoph Garbe (Interdisciplinary Center for Scientific Computing)
A tool for real-time visualization of motion features in 2D
image sequences. The motion is estimated through an eigenvector
analysis of the spatio-temporal structure tensor at every pixel
location.
Dispersion Simulation and Visualization for Urban Security
Feng Qiu, Ye Zhao, Zhe Fan, Xiaoming Wei, Haik Lorenz, Jianning Wang, Suzanne Yoakum-Stover, Arie Kaufman, Klaus Mueller (Center for Visual Computing and Department of Computer
Science, Stony Brook University)
A system for simulating and visualizing the propagation
of dispersive contaminants with an application to urban security.
VIS: Capstone Address
Self-Illustrating Phenomena
Pat Hanrahan (Stanford University)
“A self-illustrating phenomenon is an image which exposes the science behind it […]. Some famous
examples are pictures of iron filings aligned along magnetic lines of force, sand particles collecting at the
stationary points of the standing waves of a violin, stress in a mechanical part revealed through
birefringence, and particle tracks in a bubble chamber. Such images […] combine experimental
design, analysis, and visualization. […]
Computer tools for visual analysis should normally be conceived of as aids in constructing computational visual
experiments; and that the resulting visualizations be consciously designed to help validate or invalidate the
hypothesis being tested by the experiment.”
Conclusion

Difficult to see all. I made choices. Nice to have all the proceedings immediately. All is on a CD (2.36 GB of data).

InfoVis and VolVis . Symposia usually separated from the conference, that times were joined with.

I preferably went to the panels because they are no paper – They are difficult to summarize but we hear many
interesting things.

People from many fields. Visualization is a tool, rarely a final goal.

2D better than 3D ? Sense making

Interactivity. “Aptic” devices.

Centered on technique, methods and algorithms.

No discussion on “which graphics system ?” (OpenGL only).

No debate on the applications used (AVS, Iris Explorer were mentioned but that was not an issue, just tests
environments)