1. Homology Groups And Persistence Homology

2. Outline Introduction Simplicial Complex Boundary Operator Homology
Triangulation
Persistent Homology

3. Introduction Why we need homology ? Connected Components =2 Holes=20
Tunnels=1059, Cavities=0

4. Simplices 0-Simplex Point ∆0 1-Simplex Line Segment ∆1 2-Simplex
Triangle ∆2
3-Simplex
Tetrahedron ∆3
Persistence Homology

5. Simplicial Complex
A simplicial complex К is a finite collection of set of simplices that satisfies the following conditions:
Every face of a simplex of К is also in К .
The intersection of any two simplices of К is a face of each of them. .
Vertices are 0 faces, edge are 1 face etc
Add equations
К={(1,2,3)
(1,2),(1,3),(2,3),(2,4),(3,4)
(1),(2),(3),(4)}
Invalid Simplicial complex
Simplicial complex
Persistence Homology
J. R. Munkres, Elements of Algebraic Topology, p. 7,

6. Chain Complex
Let К ={σik} be a simplicial complex with simplices σik, where k denotes the simplex dimension. A simplicial k-chain is a formal sum of k-dimensional simplices
C0=A+B
C1=a+b+c
Persistence Homology

7. Boundary Operator
The boundary operator ∂, acting on simplices is a following map
Boundaries have no boundaries
Persistence Homology
Algebraic Topology, Hatcher

8. Cycles and Boundaries A chain is a cycle when its boundary is zero
The cycles form a subgroup Zk(К) of chain group Ck(К), which is the kernel of boundary operator (Z is because of german word of cycle)
Zk(К) =ker(∂k)
The elements in Im(∂k+1) are called boundaries
The k-boundary group of К is the set of boundaries of (k+1)-chains in К, i.e. Its the Image of the (k+1)-chain group
Bk(К)= ∂(Ck+1(К))
Persistence Homology

9. Finally Homology!!
The homology group is the quotient vector space of cycles over boundaries
Hk (К)= Zk(К) / Bk(К)
Suppose that V= {(x1,x2,x3)} and W= {(x1,0,0)}, then quotient vector space V/W (read as " mod") is isomorphic to {(x2,x3)}=R2
Persistence Homology
Intelligent Perception, Computer Vision Primer

10. Cycles and Boundaries 0-Simplex = {A,B,C} 0-Simplex = {A,B,C}
2-Simplex = empty
0-Simplex = {A,B,C}
1-Simplex = {a,b,c}
2-Simplex = {f}
∂2f=a+b+c
H1=Z1/B1=0
Persistence Homology

11. Homology of a Circle (S1)
Vertices v
Edges e
Boundary (∂1)
∂e=v-v
∆0(S1)
∆1(S1) H0 H1
Persistence Homology

12. Computing Homology
Cycles which generate the n dimensional holes are called homology generators
Agoston Algorithm(1976)
Build incidence matrices
Reduce to smith normal form
Compute homology Generators
Computing Homology Group Generators of Images Using Irregular Graph Pyramids, S. Peltier
Persistence Homology

13. Computing Homology
For large no. of vertices Agoston algorithm becomes computationally very expensive
Solution: Build a pyramid
It reduces no. of cells
Apply Agoston Algorithm at top level
Generators fit nicely on borders
Computing Homology Group Generators of Images Using Irregular Graph Pyramids, S. Peltier
Persistence Homology

14. The problem of assigning simplices to point cloud
Persistence Homology

15. Delauny Triangulation
For a set P of points in the plane, a triangulation DT(P) such that no point in P is inside the circumcircle of any triangle in DT(P).
The DT(P) is unique if all the points are in general position in e.g in 2-dimensional space
No three points are on same line
No four points are on same circle
Persistence Homology

16. Application (Terrain Model)
Set of data points A  R2
Height ƒ(p) defined at each point p in A
How can we most naturally approximate height of points not in A?
Persistence Homology
* Delauny Triangulations by Glenn Eguchi

17. Alpha Shapes
Given a finite point set S, and a real parameter alpha: The set of all real numbers alpha leads to a family of shapes capturing the intuitive notion of "crude'' versus "fine'' shape of a point set
Persistence Homology

18. Alpha Shapes Continued.....
For sufficiently large alpha, the alpha shape is identical to the convex hull of S
For α=0, it reduces to point cloud
Persistence Homology
Three Dimensional Alpha Shapes, Herbert Edelsbrunner

19. Filtration
A persistence complex C is a family of chain complexes C*i , together with a chain map
Persistence Homology
Computing Persistent Homology, Afra Zomorodian

20. Persistence
Persistence is a measure of importance of an n-cycle defined to be the
difference between the  for which the cycle is created, to the  it is filled by adding an (n+1)-simplex.
Persistence Homology
Persistent Homology of Complex Networks, D. Horak

21. Persistence of complex Networks
New approach to study highly interconnected dynamic systems such as scale free networks (e.g. Airline traffic routes)
Persistence of the complex gives important information about robustness of the network against addition or removal of nodes
Persistence Homology
Persistent Homology of Complex Networks, D. Horak

22. Persistence continued
Cycles which have low persistence can be regarded as topological noise
Persistence Homology
Barcodes: The Persistent Topology Of Data, Robert Ghrist

23. Persistent Homology The p-persistent k-th homology groupof Kl is
Persistence Homology
Barcodes: The Persistent Topology Of Data, Robert Ghrist
Topological persistence and Simplification, Edelsbrunner, Zomorodian

24. Protein Structure Protein function is in part determined by its shape
This shape allows it to bind to a target molecule
One important and challenging goal of proteomics, the study of proteins, is the identification and characterization of protein binding sites.
Protein data bank website contain 34,303 structures
Persistence Homology
Applications of Computational Homology, Master thesis, Marshall University

25. Protein Structure
Shapes such as letter “C” may nearly be like a circle, but not quite, so we want to capture such structures as well
Hand like structures (TAQ Polymerase) can not be perceived by just looking at the betti numbers of the structure
Amount of time a cycles is created and detroyed can give important features
Persistence Homology

26. Another application!!!
Astronomers have measured the location of about galaxies, each one represented by a point in three-dimensional space.
It appears that a large number of galaxies are located on or close to sheet-like and to lament-shaped structures.
In other words, large subsets of the points are distributed in a predominantly two- or one-dimensional manner
Persistence Homology
Topology for computing, Afra Zomorodian, p-228
M D Dyksterhouse. An alpha-shape view of our universe.

27. Summary
Homology classifies objects based on their connectivity and n-dimensional holes
Computing homology using pyramids produces nice generators and is computationally inexpensive than previous methods
Alpha shapes provide new tool in analyzing topological properties of the objects
Current research of alpha shapes and persistence homology has mostly focused on molecular biology, but its application in other fields is growing fast.
Persistence Homology

28. References
Elements of Algebraic Topology, James R Munkres, MIT, Massacheussets 1984
Algebraic topology, Allen Hatcher, Cambridge University Press, 2002
A. Zomorodian, Gunnar Carlsson, Computing Persistent Homology, Afra Zomorodian, Discrete and Computational Geometry Archive, page , Feb 2005
H. Edelsbrunner, Ernst Muecke, Three-dimensional alpha shapes, ACM Transactions on Graphics , January, 1994.
Delaunay Triangulations , Presented by Glenn Eguchi, Computational Geometry, October 11, 2001
Computer Vision Primer: beginner's guide to methods of image analysis, data analysis, related mathematics (especially topology), image analysis software, and applications in sciences and engineering.
Persistence Homology

29. References
Danijela Horak , Slobodan Maletić and Milan Rajković, Persistent Homology of Complex Networks, Institute of Nuclear Sciences Vinča, Belgrade 11001, Serbia Max Planck Institute for Mathematics in the Natural Sciences, D Leipzig, Germany, Journal of Statistical Mechanics: Theory and Experiment, Volume 2009, March 2009
H. Edelsbrunner, D. Letscher A. Zomorodian,Topological persistence and Simplification, Proceedings of the 41st Annual Symposium on Foundations of Computer Science, 2000
Barcodes: The Persistent Topology Of Data, Robert Ghrist, Department of Mathematics, University of Illinois, Urbana Champaign, 2007
Topology for computing, Afra Zomorodian, Cambridge Monographs on applied and comptational mathematics, 2005
M. D. Dyksterhouse. An alpha-shape view of our universe. Master's thesis, Department of Computer Science, University of Illinois at Urbana-Champaign, Urbana, Illinois, 1992
Applications of Computational Homology, Aaron Johnson, Master Thesis, Department of Mathematics, Marshall University, 2006
CHOMP: Computational Homology Project
Persistence Homology