Course Description
Recent advances in scanning technology and scientific
simulations can generate sample points from a geometric domain
at ease. Inferring the shape and their features from such
a simple light-weight input can be a very effective modeling paradigm
across many areas of science and engineering. In this talk
we will go over the techniques developed for such modeling
paradigms. In particular, we cover algorithms for
(i) extracting a surface out of point samples, (ii) computing
an approximate medial axis of the sampled object, (iii) segmenting
the object into so called ``features". Theoretical concepts along
with
experimental results will be presented.
Course Materials
Syllabus
- Preliminaries
- Curve Reconstruction
- Surface Samples
- Surface Reconstruction
- Undersampling
- Technique
- Results
- Noise
- Noisy samples
- Algorithm
- Noise smoothing
- MLS techniques
- AMLS
- Oversampling
- Sample Decimation
- Quality Meshing
- Large Data
- Octree subdivision
- Reconstruction
- Medial Axis Approximation
- Voronoi facet filtering
- Convergence guarantees
- Experimental results
- Shape Segmentation
- Flow and critical points
- Stable manifolds
- Discretization
- Results
- Matching
Undersampling
Oversampling
I
Oversampling
II
Large
Data
Medial
axis approximation
Shape
segmentation and matching
Dimension
detection
PMR
Papers
Surface
reconstruction
Undersampling
detection
Decimation
Large
Data
Medial
axis
Shape
segmentation
Tutorial for the Fall School at Berlin, October, 2003
Meeting
DL 266 TR 12:30-1:48
Office: DL483
Phone:292-3563
Office Hours : TR 2:00-2:30 or by appointment
Grading
Grading will be based on three things.
A. Experiment with the Cocone software
(will be provided) on data from your application
and study its results.
B. A Term paper at the end of the course on a related topic
of your choice.
C. Your class participation.