Oct. 31
Henry Fuchs
Abstract
For over a decade, we, as a Center, have shared a dream of being able to
see more than what is locally present in a fetus inside a body, a tumor
under the skin, through the wall to the other side (even if the other
side is across the country!). In the ten years we have
come a long way in learning about various aspects of technical problems
presented by this dream, and solutions toward it.
Although we've
recently had the gratification of a visualized breast biopsy on a human
patient and the illusion of looking through a portal in a wall to
another distant office, this is merely the beginning of an adventure
that will lead us into the years ahead.
The problems that remain in
data acquisition, reconstruction, transmission, presentation and
suitable merging with the surrounding world give us wonderful
opportunities for creation and intellectual growth. What was, ten years
ago, a dream is now in "proof of concept" stage, and with enthusiasm,
hard work and maybe a little luck, can become, in ten years, systems put
to daily use.
Nov. 7
David E. Breen
Abstract
The beginnings of modern cloth modeling research may be traced back
to the 1930's, with the pioneering textile science work of Peirce.
Starting in the mid-1980's the computer graphics community also began
to show great interest in modeling cloth and cloth structures for use
in computer-generated images and animation. The goals of these two
groups are very different, and therefore they each focus on different
aspects of the same problem. The textile community looks at woven cloth
as an engineering material. The computer graphics community has primarily
focused on simulating the macroscopic behavior of a single piece of cloth,
as well as a complete set of clothing, as it interacts with its environment.
The research from the graphics community can be placed in four categories,
geometric/hybrid approaches, mass-spring models, elasticity-based models,
and particle models. The first category attempts to define the geometric
structure of cloth using knowledge of the physical/mechanical behavior of
cloth, but without performing dynamic simulations. Mass-spring models
represent cloth as a network of springs and point masses. The forces produced
by the springs are applied to the masses to generate the motion of the cloth.
Elasticity-based models use simplified continuum mechanics equations and numerical
integration methods to simulate the motion of deformable sheets. Particle models
are similar to mass-spring models, except that they attempt to directly represent
the low-level structure of cloth in order to produce more accurate simulations.
This talk will summarize the research conducted on cloth modeling and animation
over the past several decades. It will primarily focus on the methods developed
for computer graphics and animation, and will highlight the evolution of these
methods since their introduction in the mid-1980's.
Nov. 21
Ross Whitaker
Abstract
Level-set methods, as proposed by Osher and Sethian (1988), can be
characterized as mechanisms for modeling the motion of curves and surfaces
using an implicit function represented on a discrete grid. Over the past
decade the technology of level sets has proved useful for solving a wide
range of problems in computational physics, computation geometry, scientific
computation, image processing, robotics, and computer vision.
This talk discusses the use of level set methods for 3D surface modeling.
The talk begins with a basic introduction to the level-set approach and the
numerical methods that are used to solve the nonlinear partial differential
equations that describe surface motion. It then gives some examples of how
this technology is used for surface-modeling problems in vision and graphics.
The talk will also discuss some of the advantages and limitations of level-set
surface modeling and some future developments for this technology.
Feb. 6
Andy van Dam
Abstract
HCI (Human-Computer Interaction) is key to effective and enjoyable use of our ever more powerful and pervasive computer and communications technology. While the quest for "faster, cheaper, better" still drives much of our industry, increasingly we've come to realize that we now have enough raw power - at least in our desktop environments - for most of what we want to do on a daily basis. This power is often unavailable, however, because we must still express our intent with clumsy, unnatural, and even crippling means of interaction. In particular, our industry is still fixated on the nearly 30-year-old WIMP interaction paradigm.
Since there is no Moore's Law for humans and since the ratio of people cost to computer cost has increased exponentially over the last few decades, it is important that we use increased computing power not just for faster information processing but to make the user interface significantly more natural and more powerful.
Our STC has not only made fundamental advances in such key areas as rendering (both photorealistic and non-photorealistic), modeling (both geometry and behavior), and scientific visualization, but has also developed novel ways for making interaction more engaging by inventing techniques to reduce the cognitive load and take better advantage of the full range of human capabilities.
In this talk, I use video to sample some of our decade-long history of work in interaction, including post-WIMP techniques, interaction in immersive environments, telecollaboration, and teleimmersion. I finish by speculating about the future of computer and interface technology and list some open research problems that must be solved before the dream of transparent and enjoyable interaction with one's computing environment can be realized.
Feb. 27
David Kirk
Abstract
The past few years have seen a revolution in performance and features for mass-market PC graphics accelerators. The first step in this revolution has been a low cost, hardwired implementation of the vanilla OpenGL and DirectX pipelines. The next frontier is programmability; graphics hardware pipelines are becoming massively programmable, allowing software developers to directly program geometry and shading operations in the graphics hardware. The innovation of programmable vertex and pixel pipelines will fundamentally change graphics. Hyper-realistic characters, special effects, and lighting and shading are now possible interactively.
Mar. 6
Zoe Wood
Abstract
We present a novel method to extract iso-surfaces from distance volumes. It generates high quality semi-regular multiresolution meshes of arbitrary topology. Our technique proceeds in two stages.
First, a very coarse mesh with guaranteed topology is extracted. Subsequently an iterative multi-scale force-based solver refines the initial mesh into a semi-regular mesh with geometrically adaptive sampling rate and good aspect ratio triangles. The coarse mesh extraction is performed using a new approach we call surface wavefront propagation. A set of discrete iso-distance ribbons are rapidly built and connected while respecting the topology of the iso-surface implied by the data.
Subsequent multi-scale refinement is driven by a simple force-based solver designed to combine good iso-surface fit and high quality sampling through reparameterization. In contrast to the Marching Cubes technique our output meshes adapt gracefully to the iso-surface geometry, have a natural multiresolution structure and good aspect ratio triangles, as demonstrated with a number of examples.
Immersive Visualization: A Ten Year Retrospective
University of North Carolina
The Evolution of Cloth Modeling and Animation
California Institute of Technology
Level-Set Surface Modeling
School of Computing, University of Utah
Interaction as Human-Centered Computing: Problems, Progress, and Prospects
Brown University
Programmability - a New Frontier in Graphics Hardware
nVidia's Chief Scientist
Semi-Regular Mesh Extraction from Volumes
Caltech
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Home | Research | Outreach | Televideo | Admin | Education |