CS 348C - Topics in Computer Graphics
Sensing for graphics
- Quarter
- Winter, 1997
- Units
- 3 (+/NC or letter grade)
- Time
- Tue/Thu 2:45 - 4:00
- Place
- 392 Gates Hall (graphics lab conference room)
- Instructors
- Marc Levoy and
Brian Curless
- Office hours
- Marc Levoy: Tue/Thu, 11:00 - 12:15
- Brian Curless: Mon/Wed, 1:15 - 2:30
- Prerequisite
- CS 248 or equivalent.
- Televised?
- No
Course abstract
This seminar course will focus on the problem of getting the physical world
into the computer. The premise is that our ability to model and animate
complex environments would be greatly enhanced if we could directly measure
properties of the physical world. Motivated by this idea and by recent
improvements in sensing technologies, the graphics community is gradually
moving toward systems that in various ways combine analytical and empirical
methods. The recent spurt in image-based modeling and rendering methods
exemplifies this trend.
As in previous versions of CS 348C, this is a mixed lecture, reading, and
project course. Lectures and readings will cover technologies and methods in
the following four areas:
- image capture:
sensors,
cameras,
digitizers,
image representations
- shape capture:
passive vrs. active sensing,
rangefinders,
surface reconstruction & fitting
- motion capture:
3D trackers,
measuring locomotion,
animation methods
- reflectance measurement:
scatterometry,
BRDF representations,
rendering methods
We will then discuss the application of these ideas to design, simulation, and
animation, and we will survey the likely impact on the manufacturing,
entertainment, games, and online markets.
Students are expected to participate in class discussions, write a short survey
paper, and do a project. A variety of
video hardware,
3D scanning and tracking systems,
and other exotic I/O devices are
available to leverage the class projects. We also have access to many
commercial modeling, animation, and rendering packages.
The course is targeted to both CS and EE students. This reflects our
conviction that successful researchers in this area must understand both the
algorithms and the underlying technologies.
levoy@cs.stanford.edu
December 12, 1996