22C:297 --- FALL 1996	

       The State of the Art in Graphics and Virtual Environments


PROFESSOR: Jim Cremer

TIME: T/Th 1:30pm

ROOM: B11 MacLean Hall


GOAL
----

In this seminar we will look at the state of the art and emerging
research directions in graphics and virtual environments (VEs).
Instead of choosing from papers from many sources over several years,
we'll focus this semester on the very latest papers.  Specifically,
we'll look at papers from the premier graphics conference, ACM
SIGGRAPH (August 1996), as well as papers from the summer 1996
editions of two well-known conferences on virtual environments, IEEE
VRAIS and ACM VRST.  I think this will make the seminar interesting in
several ways: (1) we'll make sure to get an early in-depth look at the
newest ideas and results in the field, and (2) by limiting the seminar
to these conferences, we'll be sure to spend some time looking at work
that we might not otherwise pay attention too.  Often, as our
interests become focused and specific, we don't manage to keep abreast
of interesting developments in related specialties.  So, in addition
to examining the state of the art in well-established areas or ones
we're already familiar with, I hope each participant will benefit by
learning about developments in new (either to us or to the field as a
whole) areas.

COURSE STRUCTURE, REQUIREMENTS, AND TOPICS
------------------------------------------

This will be primarily a reading and discussion course.  Each student
will be responsible for carefully reading, analyzing, and presenting
two of the conference papers to the class (see remark in NOTES below)
, with the intent to stimulate discussion.  Typically, this means that
the person will read those papers quite carefully and prepare talks,
usually using overhead transparencies, that lead the class through the
papers.  The presentations do not need to cover the full class time
--- we want to have time for discussion.  Prof. Cremer will also
present a number of papers.

In addition, students will complete a small project, either in groups
or individually, related to the course topics. For most people this
will involve some kind of programming project (using, perhaps, some of
the graphics or VE software packages we have in the research lab), but
it could also be a paper examining a problem or research area in a bit
more depth.

During the first week of class, we'll look closely at the range of
topics covered by the papers.  Topics include:

  * animation (especially ``physically-based'' and/or simulation 
      techniques)
  * geometric modeling (including collision detection)
  * image-base modeling and rendering, hierarchical rendering 
      techniques, volume rendering, rendering architectures
  * level of detail management
  * "storytelling", controlling actors/agents in virtual environments,
      scenario control
  * modeling of natural phenomena (e.g. clouds, snow, plants ...)
  * haptics, force feedback, and other VE immersion technology
  * medical and other applications of VE technology
  * software architectures for distributed and/or real-time virtual 
      environments

Certain topics might be deleted and/or additional ones added,
depending on class interests.


PREREQUISITES --- SUGGESTED BACKGROUND
--------------------------------------

There are no firm prerequisites other than interest in the material
and willingness to read some papers and participate in discussions of
research ideas.  A basic graphics course or graphics experience would
certainly help.  Familiarity with basic linear algebra, calculus,
numerical analysis, elementary physics, and so on will also be
helpful.  Please see me if you are concerned about your background.


NOTES
-----

Some of you might be nervous about presenting papers.  Consider this a
good opportunity to practice organizing and presenting a talk in a
friendly, casual environment.  This will serve you well when it comes
time to present you M.S or Ph.D final exam or thesis, or your own
conference papers.  I only require that people make a sincere effort
-- the presentations need not be highly refined or polished.