CS 284: CAGD
Lecture #1 -- Tu 8/29, 2006.
CS
284 HOME < - - - - > CURRENT
< - - - - > NEXT
Preparation:
read Rockwood: pp 1-30.
Lecture Topics
Introduction to CAGD
-
What is CAGD ?
Subset of CAD; originally mostly shipbuilding, automotive, aeroplanes;
mostly smooth, curved shapes.
-
Motivation for COE Students:
CS: smooth motions for movies; ME: freeform parts; Math: data fitting
...
- Course Content:
see information on home page... (and below)
Key Course Topics
-
Interpolation
-
Subdivision (more than in the past)
-
Splines (less than in the past)
... all on 1, 2, (and 3)-manifolds
-
Curve and surface optimization
-
T-Splines
-
Point Clouds
-
Implicit Surfaces
--- content will be tailored somewhat to interests of course participants. -
Background Questionnaire:
Please fill in !
Course Goal
- Learn how to make smooth curves and surfaces that fit certain constraints.
- TREND: Make it easier on the user (even though underlying formulation may get
more complex).
- TREND: Optimization will play an ever bigger role (compute power is available).
- A common key concept:
Use a simple (piecewise linear) shape, with the fewest degrees of
freedom (DoF) necessary,
to define and control an associated (smooth, continuous) design
shape.
Course Administration
- Where does CS 284 fit in ?
Should have had some computer graphics or CAD course;
(e.g. cs 184).
- Teaching Style:
Application oriented; learning by doing;
(even the chosen textbook has an interactive component).
Intuitive introduction first; hands-on experience; math later.
Interactive lectures, Q&A;
Many small homework assignments initially, to deepen understanding, "close feedback loop".
- The Key Problem of CAGD:
Handout: Experiment -- connect the dots ...
What do the result tell us ?
- Reading Assignments:
Important; --> you get much more out of class discussions.
-
Thinking/Experimenting/Programming Assignments:
Equally important; --> to make sure you can apply the discussed material.
-
Quizzes/Exams:
Probably just one quiz, somewhere in the middle of the semester,
to check whether the key concepts have sunk in.
-
Course
Projects:
(sorry no links -- just titles)
Reimplement the key algorithm from some (siggraph?) paper;
Create some utility that solves a particular task (possibly for your own use);
Design an interesting virtual or real object (possibly build it on our FDM machine).
Expected Math Foundations
See Rockwood pp 9-19
-
Parametric Curves and Surfaces
-- not just single valued height fields, can handle infinite slope, can be subjected to transformations ...
- Derivatives; Tangents; Normals
-- "velocity vector, vector of all component derivatives; -- if <>0, normalize to 1; -- normal on all tangent vectors
-
Linear Interpolation
-
Basis Vectors
Some Possibly New Concepts
See Rockwood pp 20-30
-
Hodograph
-- plot of parameterized derivative vector
-
Winding Number of a (closed, oriented) Curve around a Point
-- how many times does it loop around that point ?
-
Turning Number of a (closed) Curve
-- if v<>0 it is equal to winding numper of hodograph around origin.
- Parametric Continuity
-- all component functions are differentiable
-
Geometric Continuity
-- visual appearance is smooth, ignore parametrization!
-
Basis Functions
-- see definition 14, page 29 ...
New Homework Assignment:
A Conceptual Task: (to make you conscious of what we are up against).
Think about how you would model
-- with the tools that you have already at your disposal --
a completely
smooth Genus-2 Object ( = Two-hole torus).
DUE: WED 8/31/06, 2:10pm.
You don't actually have to build a CAD model for this shape,
but think through all the steps that you would have to go through,
and estimate how long it might take you.
In a couple of paragraphs, write down your thoughts on how you would
do this
and bring that write-up to class on Wednesday;
also be prepared to explain your approach in a few sentences.
Next Reading Assignment:
"RC" = Rockwood & Chambers
skim: RC pp 1-19;
review: RC pp 20-30;
read in preparation for next time: RC pp 31-48.
CS
284 HOME < - - - - > CURRENT
< - - - - > NEXT
Page Editor: Carlo H. Séquin