Computational Topology for Computer Graphics - PowerPoint PPT Presentation

1 / 50
About This Presentation
Title:

Computational Topology for Computer Graphics

Description:

Any shape is homotopic to any of its deformation retracts ... The height function f (p) = z is not good enough not rotational invariant. ... – PowerPoint PPT presentation

Number of Views:363
Avg rating:3.0/5.0
Slides: 51
Provided by: csta3
Category:

less

Transcript and Presenter's Notes

Title: Computational Topology for Computer Graphics


1
Computational Topology for Computer Graphics
Klein bottle
2
Talk Outline
  • Basic terms in Topology
  • Algebraic Topology
  • Morse Theory
  • Reeb Graph and Application

3
Talk Outline
  • Basic terms in Topology
  • Algebraic Topology
  • Morse Theory
  • Reeb Graph and Application

4
What is Topology?
  • The topology of a space is a collection of sets
    that include
  • the space and the empty set
  • the union of any of the sets
  • the finite intersection of any of the sets
  • Topological space is a set with the least
    structure necessary to define the concepts of
    nearness and continuity

5
No, Really - What is Topology?
  • The study of shape equivalence classes (under
    deformation).
  • Rules of deformation (homeomorphism)
  • Onto (all of A ? all of B)
  • 1-1 correspondence (no overlap)
  • Bicontinuous (Cant tear, join, poke/seal holes)

6
Why Topology?
  • What is the boundary of an object?
  • Are there holes in the object?
  • Is the object hollow?
  • If the object is transformed in some way, are the
    changes continuous or abrupt?
  • Is the object bounded, or does it extend
    infinitely far?

7
Topology in CG
  • The study of connectedness
  • Understanding
  • How connectivity happens?
  • Analysis
  • How to determine connectivity?
  • Articulation
  • How to describe connectivity?
  • Control
  • How to enforce connectivity?

8
For Example
  • How does connectedness affect
  • Morphing
  • Texturing
  • Compression
  • Simplification

9
Problem Mesh Reconstruction
  • Determines shape from point samples
  • Different coordinates, different shapes

10
Topological Properties
  • To uniquely determine homeomorphism we need to
    find out
  • Surface is open or closed
  • Surface is orientable or not
  • Genus (number of holes)
  • Boundary components

11
Surfaces
  • How to define surface?
  • The set of zeroes of a polynomial equation in R3
    is a 2D surface x2y2z21
  • A surface is a geometric space which locally
    looks like a plane!

12
General Surfaces Manifolds
  • An n-manifold is a topological space that
    locally looks like the  Rn plane.
  • A circle is a 1-manifold
  • A torus is a 2-manifold

13
Open vs. Closed Surfaces
  • The points having a neighborhood homeomorphic to
    R2 form Int(S) (interior)
  • The points for which every neighborhood is
    homeomorphic to R2?0 form ?S (boundary)
  • A surface S is said to be closed if its boundary
    is empty

14
Orientability
  • A surface in R3 is called orientable, if it is
    possible to distinguish between its two sides
    (inside/outside above/below)
  • A non-orientable surface has a path which brings
    a traveler back to his starting point
    mirror-reversed (inverse normal)

15
Orientation by Triangulation
  • A surface is orientable if it has a consistent
    CW/CCW normal assignment to its triangulation
  • If any two triangles sharing an edge have
    opposite directions on that edge the surface is
    orientable.

16
Genus and holes
  • Genus of a surface is the maximal number of
    nonintersecting simple closed curves that can be
    drawn on the surface without separating it
  • The genus is equivalent to the number of holes or
    handles on the surface
  • Example
  • Genus 0 point, line, sphere
  • Genus 1 torus
  • Genus 2 the symbol 8

17
Euler characteristic function
  • Polyhedral decomposition of a surface
  • (V vertices, E edges, F faces)
  • ?(M) V E F
  • If M has g holes and h boundary components
    then ?(M) 2 2g h

? 1
18
Homeomorphism in R3
  • Any orientable closed surface is topologically
    equivalent to a sphere with g handles attached to
    it

19
Talk Outline
  • Basic terms in Topology
  • Algebraic Topology
  • Morse Theory
  • Reeb Graph and Application

20
Algebraic Topology
  • Homeomorphism gives much too fine a
    classification to be useful
  • Algebraic Topology
  • Homotopy equivalence
  • Deformation retraction
  • Cells

21
Homotopy equivalence
  • A B ? There is a continuous map between A and
    B
  • Same number of components
  • Same number of holes
  • Not necessarily the same dimension

22
Deformation Retraction
  • Function that continuously reduces a set
    onto a subset
  • Any shape is homotopic to any of its deformation
    retracts
  • Skeleton is a deformation retract of the solids
    it defines

23
Cells
  • Cells are dimensional primitives

0-cell
1-cell
2-cell
3-cell
24
Talk Outline
  • Basic terms in Topology
  • Algebraic Topology
  • Morse Theory
  • Reeb Graph and Application

25
Hard Problems Dunking a Donut
  • Dunk the donut in the coffee!
  • Investigate the change in topology of the portion
    of the donut immersed in the coffee

26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Solution Morse Theory
  • Investigates the topology of a surface by the
    critical points of a real function on the surface
  • Critical point occur where the gradient ??f
    (??f/?x, ?f/?y,) 0
  • Index of a critical point is of principal
    directions where f decreases

32
Example Dunking a Donut
  • Surface is a torus
  • Function f is height
  • Investigate topology of f ?? h
  • Four critical points
  • Index 2 maximum
  • Index 1 saddle
  • Index 1 saddle
  • Index 0 minimum

33
Morse function
  • f doesnt have to be height any Morse function
    would do
  • f is a Morse function on M if
  • f is smooth
  • All critical points are isolated
  • Critical points are non-degenerate
  • det(Hessian(p)) ! 0

34
Critical Point Index
  • The index of a critical point is the number of
    negative eigenvalues of the Hessian
  • 2 ? maximum
  • 1 ? saddle point
  • 0 ? minimum
  • Intuition the number of independent
    directions in which f
    decreases

35
If sweep doesnt pass critical pointMilnor 1963
  • Denote Ma p ? M f(p) ? a (the sweep region
    up to value a of f )
  • If f ?1a, b is compact and doesnt contain
    critical points of f Ma is homeomorphic to Mb.

36
Sweep passes critical pointMilnor 1963
  • If p is critical point of f with index ?, then
    Mc? has the same homotopy type as Mc?? with
    ?-cell attached.

Mc?
Mc??

Mc?
37
This is what happened to the doughnut
38
Conclusions
  • Topology describes properties of shape that are
    invariant under deformations
  • Morse theory connects between differential
    geometry and topology of surface
  • We can investigate topology by investigating
    critical points of Morse functions
  • Looking at topology of level sets (sweeps) of a
    Morse function, we can learn about its critical
    points.

39
Talk Outline
  • Basic terms in Topology
  • Algebraic Topology
  • Morse Theory
  • Reeb Graph and Application

40
Reeb graphs
  • Schematic way to present a Morse function
  • Vertices of the graph are critical points
  • Arcs of the graph are contracted connected
    components of the level sets of f,

2
1
1
1
1
1
0
0
41
Reeb graphs and genus
  • The number of loops in the Reeb graph is equal to
    the surface genus
  • To count the loops, simplify the graph by
    contracting degree-1 vertices and removing
    degree-2 vertices

degree-1
degree-2
42
Another Reeb graph example
43
Reeb graphs for Shape Matching
  • Reeb graph encodes the behavior of a Morse
    function on the shape
  • Also tells us about the topology of the shape
  • Take a meaningful function and use its Reeb graph
    to compare between shapes!

44
Choose the right Morse function
  • The height function f (p) z is not good enough
    not rotational invariant.
  • Not always a Morse function

45
Average geodesic distance
  • The idea of Hilaga et al. 01 use geodesic
    distance for the Morse function!

46
Multi-res Reeb graphs
  • Hilaga et al. use multiresolutional Reeb graphs
    to compare between shapes
  • Multiresolution hierarchy by gradual
    contraction of vertices

47
The End!
48
Mesh Partitioning
  • Now we get to Zhang et al. 03
  • They use almost the same f as Hilaga et al. 01
  • Want to find features long protrusions
  • Find local maxima of f !

49
Region growing
  • Start the sweep from global minimum (central
    point of the shape)
  • Add one triangle at a time the one with
    smallest f
  • Record topology changes in the boundary of the
    sweep front these are critical points

50
Critical points genus-0 surface
  • Splitting saddle when the front splits into two
  • Maximum when one front boundary component
    vanishes

min
splitting saddle
max
max
Write a Comment
User Comments (0)
About PowerShow.com