Reeb Graphs and Piecewise Linear Functions Koen Klaren Eindhoven - - PowerPoint PPT Presentation

reeb graphs and piecewise linear functions
SMART_READER_LITE
LIVE PREVIEW

Reeb Graphs and Piecewise Linear Functions Koen Klaren Eindhoven - - PowerPoint PPT Presentation

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Reeb Graphs and Piecewise Linear Functions Koen Klaren Eindhoven University of Technology Seminar Algorithms, 2018 Reeb graphs Piecewise linear functions Constructing


slide-1
SLIDE 1

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary

Reeb Graphs and Piecewise Linear Functions

Koen Klaren

Eindhoven University of Technology

Seminar Algorithms, 2018

slide-2
SLIDE 2

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary

Outline

1 Reeb graphs 2 Piecewise linear functions 3 Constructing Reeb graphs

slide-3
SLIDE 3

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Outline

1 Reeb graphs

Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

2 Piecewise linear functions 3 Constructing Reeb graphs

slide-4
SLIDE 4

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Motivation for Reeb graphs

Make the structure of a function explicit by visualizing the evolution of the components of the level set

Reeb graph

slide-5
SLIDE 5

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Motivation for Reeb graphs

Make the structure of a function explicit by visualizing the evolution of the components of the level set Accelerate extraction of isosurfaces (level sets)

Reeb graph Isosurface extraction

slide-6
SLIDE 6

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Motivation for Reeb graphs

Make the structure of a function explicit by visualizing the evolution of the components of the level set Accelerate extraction of isosurfaces (level sets) Applications in medical imaging and other areas of science and engineering

Reeb graph Isosurface extraction

slide-7
SLIDE 7

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Outline

1 Reeb graphs

Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

2 Piecewise linear functions 3 Constructing Reeb graphs

slide-8
SLIDE 8

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

Given a continuous map f : X → R, the level sets form a partition

  • f topological space X
slide-9
SLIDE 9

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

slide-10
SLIDE 10

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

Given a continuous map f : X → R, the level sets form a partition

  • f topological space X

Definition Two points x, y ∈ X are called equivalent if they belong to a common component in a level set of f

slide-11
SLIDE 11

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

slide-12
SLIDE 12

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

Given a continuous map f : X → R, the level sets form a partition

  • f topological space X

Definition Two points x, y ∈ X are called equivalent if they belong to a common component in a level set of f Definition Defined equivalence classes are the contours of f

slide-13
SLIDE 13

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of contours

slide-14
SLIDE 14

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

Definition The Reeb graph of f is the set of contours R(f).

slide-15
SLIDE 15

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

Definition The Reeb graph of f is the set of contours R(f).

slide-16
SLIDE 16

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

We started with only the continuous map f. Where does the Reeb graph fit in? X R

f ψ π

slide-17
SLIDE 17

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

Let ψ: X → R(f) be the function that maps a point x to the contour ψ(x) that contains it: X R R(f)

f ψ π

slide-18
SLIDE 18

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

Let π: R(f) → R be the unique map whose composition with ψ is f: X R R(f)

f ψ π

π(ψ(x)) = f(x)

slide-19
SLIDE 19

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Definition of Reeb graphs

This speeds up construction of a level set, f−1(a). We can use π−1(a) to find contours r ∈ R(f) that form the level set. X R R(f)

f ψ π

slide-20
SLIDE 20

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Outline

1 Reeb graphs

Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

2 Piecewise linear functions 3 Constructing Reeb graphs

slide-21
SLIDE 21

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

Connected components

Function ψ: X → R(f) maps components to components

slide-22
SLIDE 22

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

Connected components

Function ψ: X → R(f) maps components to components Expressed in terms of Betti numbers, we get: β0(R(f)) = β0(X)

slide-23
SLIDE 23

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

1-dimensional connectivity

In some cases the Reeb graph reflects the 1-dimensional connectivity of the space

slide-24
SLIDE 24

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

1-dimensional connectivity

In some cases the Reeb graph reflects the 1-dimensional connectivity of the space A 1-cycle in R(f) is called a loop The preimage of a loop is non-contractible in X Different loops correspond to non-homologous 1-cycles in X

slide-25
SLIDE 25

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

1-dimensional connectivity

In some cases the Reeb graph reflects the 1-dimensional connectivity of the space A 1-cycle in R(f) is called a loop The preimage of a loop is non-contractible in X Different loops correspond to non-homologous 1-cycles in X Expressed in terms of Betti numbers, we get: β1(R(f)) ≤ β1(X)

slide-26
SLIDE 26

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Properties of Reeb graphs

Contour tree

From these properties: β0(R(f)) = β0(X) β1(R(f)) ≤ β1(X) it follows that if X is contractible then R(f) is a tree, independent of the function f Because of this R(f) is sometimes called a contour tree

slide-27
SLIDE 27

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Outline

1 Reeb graphs

Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

2 Piecewise linear functions 3 Constructing Reeb graphs

slide-28
SLIDE 28

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Reeb graphs of Morse functions

More can be said if X = M is a manifold of dimension d ≥ 2 and f : M → R is a Morse function

slide-29
SLIDE 29

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Reeb graphs of Morse functions

More can be said if X = M is a manifold of dimension d ≥ 2 and f : M → R is a Morse function Recall that each point r ∈ R(f) is the image of a contour in M Call r a node if it is the image of a critical point under ψ The other points form arcs that connect the nodes

slide-30
SLIDE 30

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Reeb graphs of Morse functions

slide-31
SLIDE 31

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Reeb graphs of Morse functions

More can be said if X = M is a manifold of dimension d ≥ 2 and f : M → R is a Morse function Recall that each point r ∈ R(f) is the image of a contour in M Call r a node if it is the image of a critical point under ψ The other points form arcs that connect the nodes By definition of Morse functions, this implies a bijection between the critical points of f and the nodes of R(f)

slide-32
SLIDE 32

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

Minimum starts a contour and thus corresponds to a degree

  • ne node
slide-33
SLIDE 33

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

slide-34
SLIDE 34

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

Minimum starts a contour and thus corresponds to a degree

  • ne node

Index 1 saddle that merges two contours into one corresponds to a degree three node

slide-35
SLIDE 35

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

slide-36
SLIDE 36

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

Minimum starts a contour and thus corresponds to a degree

  • ne node

Index 1 saddle that merges two contours into one corresponds to a degree three node Index d − 1 saddle that splits a contour into two corresponds to a degree three node

slide-37
SLIDE 37

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

slide-38
SLIDE 38

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

Minimum starts a contour and thus corresponds to a degree

  • ne node

Index 1 saddle that merges two contours into one corresponds to a degree three node Index d − 1 saddle that splits a contour into two corresponds to a degree three node Maximum closes a contour and thus corresponds to a degree

  • ne node
slide-39
SLIDE 39

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

slide-40
SLIDE 40

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definitions Properties of Reeb graphs Reeb graphs of Morse functions

Bijection between critical points and nodes

Minimum starts a contour and thus corresponds to a degree

  • ne node

Index 1 saddle that merges two contours into one corresponds to a degree three node Index d − 1 saddle that splits a contour into two corresponds to a degree three node Maximum closes a contour and thus corresponds to a degree

  • ne node

All other critical points correspond to nodes of degree two

slide-41
SLIDE 41

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Outline

1 Reeb graphs 2 Piecewise linear functions

Motivation Definition of PL functions Lower star filtration PL Morse functions

3 Constructing Reeb graphs

slide-42
SLIDE 42

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Motivation for piecewise linear (PL) functions

Functions are rarely smooth in practical situations

slide-43
SLIDE 43

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Motivation for piecewise linear (PL) functions

Functions are rarely smooth in practical situations Instead we have non-smooth functions that approximate smooth ones

slide-44
SLIDE 44

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Motivation for piecewise linear (PL) functions

Functions are rarely smooth in practical situations Instead we have non-smooth functions that approximate smooth ones Thus we introduce piecewise linear (PL) functions

slide-45
SLIDE 45

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Outline

1 Reeb graphs 2 Piecewise linear functions

Motivation Definition of PL functions Lower star filtration PL Morse functions

3 Constructing Reeb graphs

slide-46
SLIDE 46

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Definition of PL functions

Let K be a simplicial complex with real values specified at all

  • vertices. Using linear extension over the simplices, we obtain a

piecewise linear (PL) function f : |K| → R.

slide-47
SLIDE 47

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Genericity

Assume that f is generic, meaning that all vertices have distinct function values

slide-48
SLIDE 48

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Genericity

Assume that f is generic, meaning that all vertices have distinct function values Now the vertices can be ordered by increasing function value: f(u1) < f(u2) < . . . < f(un)

slide-49
SLIDE 49

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Genericity

Assume that f is generic, meaning that all vertices have distinct function values Now the vertices can be ordered by increasing function value: f(u1) < f(u2) < . . . < f(un) Let Ki be the full subcomplex defined by the first i vertices

slide-50
SLIDE 50

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Outline

1 Reeb graphs 2 Piecewise linear functions

Motivation Definition of PL functions Lower star filtration PL Morse functions

3 Constructing Reeb graphs

slide-51
SLIDE 51

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Lower star filtration

Recall that the star of a vertex ui is the set of cofaces of ui in K Definition The lower star is the subset of simplices in the star of ui for which ui is the vertex with maximum function value: St ui = {σ ∈ St ui | x ∈ σ ⇒ f(x) ≤ f(ui)}

slide-52
SLIDE 52

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Lower star filtration

By assumption of genericity, each simplex has a unique maximum vertex and thus belongs to a unique lower star

slide-53
SLIDE 53

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Lower star filtration

By assumption of genericity, each simplex has a unique maximum vertex and thus belongs to a unique lower star It follows that the lower stars partition K. Also, Ki is the union of the first i lower stars.

slide-54
SLIDE 54

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Lower star filtration

By assumption of genericity, each simplex has a unique maximum vertex and thus belongs to a unique lower star It follows that the lower stars partition K. Also, Ki is the union of the first i lower stars. This nested sequence of subcomplexes is thus called the lower star filtration of f: ∅ = K0 ⊆ K1 ⊆ . . . ⊆ Kn = K

slide-55
SLIDE 55

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Outline

1 Reeb graphs 2 Piecewise linear functions

Motivation Definition of PL functions Lower star filtration PL Morse functions

3 Constructing Reeb graphs

slide-56
SLIDE 56

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

Lower link

Recall that the link of a vertex ui is the set of simplices in the closed star that do not belong to the star Definition The lower link is the set of simplices in the closed lower star that do not belong to the lower star: Lk ui = {σ ∈ Lk ui | x ∈ σ ⇒ f(x) < f(ui)}

slide-57
SLIDE 57

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

As we go from Ki−1 to Ki we attach the closed lower star of ui, gluing it along the lower link to the complex Ki−1

slide-58
SLIDE 58

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

As we go from Ki−1 to Ki we attach the closed lower star of ui, gluing it along the lower link to the complex Ki−1 Homology of the lower link is used to classify vertices

slide-59
SLIDE 59

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

Regular

Call ui a PL regular vertex if its lower link is non-empty but homologically trivial

slide-60
SLIDE 60

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

Regular Critical (q = 0)

Call ui a PL regular vertex if its lower link is non-empty but homologically trivial Call ui a simple PL critical vertex of index q if its lower link has the reduced homology of the (q − 1)-sphere

slide-61
SLIDE 61

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

Regular Critical (q = 0) Critical (q = 1)

Call ui a PL regular vertex if its lower link is non-empty but homologically trivial Call ui a simple PL critical vertex of index q if its lower link has the reduced homology of the (q − 1)-sphere

slide-62
SLIDE 62

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL critical points

Regular Critical (q = 0) Critical (q = 1) Critical (q = 2)

Call ui a PL regular vertex if its lower link is non-empty but homologically trivial Call ui a simple PL critical vertex of index q if its lower link has the reduced homology of the (q − 1)-sphere

slide-63
SLIDE 63

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Motivation Definition of PL functions Lower star filtration PL Morse functions

PL Morse functions

Definition We call a PL function f : |K| → R on a manifold a PL Morse function if:

1 Each vertex is either PL regular or simple PL critical 2 The function values of the vertices are distinct

slide-64
SLIDE 64

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Outline

1 Reeb graphs 2 Piecewise linear functions 3 Constructing Reeb graphs

Sketch of the algorithm Running time

slide-65
SLIDE 65

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Constructing a Reeb graph

Problem: Construct a Reeb graph of a function on a 2-manifold. We assume M is triangulated and the function f : M → R is PL Morse.

slide-66
SLIDE 66

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Constructing a Reeb graph

Problem: Construct a Reeb graph of a function on a 2-manifold. We assume M is triangulated and the function f : M → R is PL Morse. Remember we can order the vertices of M: f(u1) < f(u2) < . . . < f(un)

slide-67
SLIDE 67

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Constructing a Reeb graph

Problem: Construct a Reeb graph of a function on a 2-manifold. We assume M is triangulated and the function f : M → R is PL Morse. Remember we can order the vertices of M: f(u1) < f(u2) < . . . < f(un) We can use a plane sweep algorithm!

slide-68
SLIDE 68

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Constructing a Reeb graph

Problem: Construct a Reeb graph of a function on a 2-manifold. We assume M is triangulated and the function f : M → R is PL Morse. Remember we can order the vertices of M: f(u1) < f(u2) < . . . < f(un) We can use a plane sweep algorithm! Status is the set of contours at the current level Events take place at each of the vertices

slide-69
SLIDE 69

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

Four cases:

1 ui is a minimum: add a degree one node to the Reeb graph.

Start a new arc associated with the new contour.

slide-70
SLIDE 70

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

slide-71
SLIDE 71

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

Four cases:

1 ui is a minimum: add a degree one node to the Reeb graph.

Start a new arc associated with the new contour.

2 ui is a regular vertex: update the associated contour in the

status.

slide-72
SLIDE 72

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

slide-73
SLIDE 73

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

Four cases:

1 ui is a minimum: add a degree one node to the Reeb graph.

Start a new arc associated with the new contour.

2 ui is a regular vertex: update the associated contour in the

status.

3 ui is a saddle: three different cases depending on the index.

Usually merge/split contours in the status and add a degree three node to the Reeb graph.

slide-74
SLIDE 74

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

slide-75
SLIDE 75

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

Four cases:

1 ui is a minimum: add a degree one node to the Reeb graph.

Start a new arc associated with the new contour.

2 ui is a regular vertex: update the associated contour in the

status.

3 ui is a saddle: three different cases depending on the index.

Usually merge/split contours in the status and add a degree three node to the Reeb graph.

4 ui is a maximum: Remove contour from the status and add a

degree one node to the Reeb graph.

slide-76
SLIDE 76

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Event handling

slide-77
SLIDE 77

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Outline

1 Reeb graphs 2 Piecewise linear functions 3 Constructing Reeb graphs

Sketch of the algorithm Running time

slide-78
SLIDE 78

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Running time

Implement the contour status using cyclic lists in a balanced search tree

slide-79
SLIDE 79

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Running time

Implement the contour status using cyclic lists in a balanced search tree This data structure supports all required operations in O(log2 m) time, where m is the number of edges in M

slide-80
SLIDE 80

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary Sketch of the algorithm Running time

Running time

Implement the contour status using cyclic lists in a balanced search tree This data structure supports all required operations in O(log2 m) time, where m is the number of edges in M Thus the algorithm constructs a Reeb graph in O(m log2 m) time

slide-81
SLIDE 81

Reeb graphs Piecewise linear functions Constructing Reeb graphs Summary

Summary

Reeb graphs as a data structure for the extraction of level sets and as a tool to analyse topological spaces Piecewise linear functions as a more practical alternative to smooth functions Sketched an efficient algorithm for construction of Reeb graphs for 2-manifolds