CS 284: CAGD
Lecture #20 -- Mo 11/03, 2003.
PREVIOUS
< - - - - > CS
284 HOME < - - - - > CURRENT
< - - - - > NEXT
Preparation:
READ: "Implicit
Modeling using Subdivivision-Curves" S. Hornus et al.
It is also helpful to read first:
"Adaptive Sampling
of Implicit Surfaces for Interactive Modeling and Animation" by M.
Desbrun et al.
Modeling with Implicit Surfaces and Conversion to B-Reps
Presentation and discussion led by Alex Krasnov:
"Implicit
Modeling using Subdiv.Curves" Hornus et al.
Some self-test questions:
-
What does the user input to this design system ?
-
What determines the number of sampling points on the surface ?
-
Is the field around two collinear segments, joined at their ends, the same
as that of a single line segment equal in lenght to their combined
length ?
-
How do they avoid unwanted blending between unrelated bodies (segments)
?
-
What are the options for visualizing these surfaces ?
Issues to ponder:
-
What kind of continuity do these implicit surfaces achieve ?
-
Can we put a C1-continuous texture onto these surfaces ?
Comments on Edge-based Surface Energies
Warm-up Exercise: Bending
energy of poly-lines:
-
In-circle model (gives too much curvature):
-
Circum-circle model (gives too little curvature):
-
Using the angle itself should be about right in the limit:
Extend this model to: Polyhedral
surfaces:
-
Assume all the bending energy is associated with the edges (seems very
reasonable),
-
This energy is representing the area from the edge itself to the centroids
of the faces on either side.
Approximating a sphere with the Platonic solids and with some semi-regular
solids:
-
The edge-based bending energy approximates the total bending energy of
the sphere (W = 4p = 12.57)
the better, the more edges there are, and the more regular the tessellation
is:
-
Some
sample results:
Preparing for 5-Minute Project Proposal Presentations (11/5/03)
("Venture Capitalist Rally" -- How to get the "gold".)
Some
hints on how to make a good presentation.
Your peers will rank-order your performance: A
glipse at the score card:
Let's start at 9:00am sharp, so we have a little
less time pressure:
-
Should last 4-5 minutes (hard 5 minute cutoff).
-
Need to switch from one speaker to next one in 1 minute -- be prepared
for this !
-
Start your talk with an attention-grabber.
-
Be enthusiastic, leave a (positive) impression.
-
Show relevance, potential impact of your work.
-
Be on time ! (you may be first at 9:00am !)
Next Reading Assignments: (see
list)
Optimize your genus-4 object with the Brakke Surface Evolver.
If you rather optimize a different object of similar complexity --
come talk to me.
PREVIOUS
< - - - - > CS
284 HOME < - - - - > CURRENT
< - - - - > NEXT
Page Editor: Carlo H. Séquin