1 Euler Number Computation By Kishore Kulkarni (Under the Guidance of Dr. Longin Jan Latecki)

Slides:



Advertisements
Similar presentations
Goal: a graph representation of the topology of a gray scale image. The graph represents the hierarchy of the lower and upper level sets of the gray level.
Advertisements

Computer Graphics- SCC 342
Image Segmentation Longin Jan Latecki CIS 601. Image Segmentation Segmentation divides an image into its constituent regions or objects. Segmentation.
UNC Chapel Hill Lin/Foskey/Manocha Steps in DP: Step 1 Think what decision is the “last piece in the puzzle” –Where to place the outermost parentheses.
Lecture 5 Graph Theory. Graphs Graphs are the most useful model with computer science such as logical design, formal languages, communication network,
MCA 202: Discrete Mathematics under construction Instructor Neelima Gupta
Divide and Conquer. Subject Series-Parallel Digraphs Planarity testing.
De Bruijn sequences Rotating drum problem:
Midwestern State University Department of Computer Science Dr. Ranette Halverson CMPS 2433 – CHAPTER 4 GRAPHS 1.
Map Overlay Algorithm. Birch forest Wolves Map 1: Vegetation Map 2: Animals.
2/14/13CMPS 3120 Computational Geometry1 CMPS 3120: Computational Geometry Spring 2013 Planar Subdivisions and Point Location Carola Wenk Based on: Computational.
Matthew L. WrightMatthew L. Wright Institute for Mathematics and its ApplicationsInstitute for Mathematics and its Applications University of MinnesotaUniversity.
Graph Search Methods Spring 2007 CSE, POSTECH. Graph Search Methods A vertex u is reachable from vertex v iff there is a path from v to u. A search method.
Extended Gaussian Images
Tirgul 8 Graph algorithms: Strongly connected components.
DIGITAL IMAGE PROCESSING
Foreground/Background Image Segmentation. What is our goal? To label each pixel in an image as belonging to either the foreground of the scene or the.
Digital Image Processing Lecture 12: Image Topology
Image Topology ( Part 3 ) Dr. Bai-ling Zhang School of Computer Science & Mathematics Victoria University of Technology SCM3511 Image Processing 2, Week.
Steps in DP: Step 1 Think what decision is the “last piece in the puzzle” –Where to place the outermost parentheses in a matrix chain multiplication (A.
3. Delaunay triangulation
Morphology Structural processing of images Image Processing and Computer Vision: 33 Morphological Transformations Set theoretic methods of extracting.
Connected Components, Directed Graphs, Topological Sort COMP171.
Processing Digital Images. Filtering Analysis –Recognition Transmission.
Tucker, Applied Combinatorics, Section 1.4, prepared by Patti Bodkin
MRF Labeling With Graph Cut CMPUT 615 Nilanjan Ray.
Drawing of G. Planar Embedding of G Proposition Proof. 1. Consider a drawing of K 5 or K 3,3 in the plane. Let C be a spanning cycle. 2. If the.
November 29, 2004AI: Chapter 24: Perception1 Artificial Intelligence Chapter 24: Perception Michael Scherger Department of Computer Science Kent State.
CSE554Binary PicturesSlide 1 CSE 554 Lecture 1: Binary Pictures Fall 2013.
CSE554Binary PicturesSlide 1 CSE 554 Lecture 1: Binary Pictures Fall 2014.
Copyright © Cengage Learning. All rights reserved. CHAPTER 11 ANALYSIS OF ALGORITHM EFFICIENCY ANALYSIS OF ALGORITHM EFFICIENCY.
Edge Detection Hao Huy Tran Computer Graphics and Image Processing CIS 581 – Fall 2002 Professor: Dr. Longin Jan Latecki.
Chapter 3 Binary Image Analysis. Types of images ► Digital image = I[r][c] is discrete for I, r, and c.  B[r][c] = binary image - range of I is in {0,1}
Lecture 5. Morphological Image Processing. 10/6/20152 Introduction ► ► Morphology: a branch of biology that deals with the form and structure of animals.
CS 6825: Binary Image Processing – binary blob metrics
An arrangement of lines: A(H)
Copyright © Cengage Learning. All rights reserved. CHAPTER 10 GRAPHS AND TREES.
Straight line drawings of planar graphs – part II Roeland Luitwieler.
Discrete Math for CS Binary Relation: A binary relation between sets A and B is a subset of the Cartesian Product A x B. If A = B we say that the relation.
12. Polyhedra
Introduction Image geometry studies rotation, translation, scaling, distortion, etc. Image topology studies, e.g., (i) the number of occurrences.
Unit – V Graph theory. Representation of Graphs Graph G (V, E,  ) V Set of vertices ESet of edges  Function that assigns vertices {v, w} to each edge.
Discrete Mathematical Structures: Theory and Applications
Image Segmentation by Histogram Thresholding Venugopal Rajagopal CIS 581 Instructor: Longin Jan Latecki.
Morphological Image Processing การทำงานกับรูปภาพด้วยวิธีมอร์โฟโลจิคัล
CS654: Digital Image Analysis
12.1 Exploring Solids.
MAT 2720 Discrete Mathematics Section 8.2 Paths and Cycles
Lecture 17: Trees and Networks I Discrete Mathematical Structures: Theory and Applications.
8.4 Closures of Relations Definition: The closure of a relation R with respect to property P is the relation obtained by adding the minimum number of.
Image Segmentation Nitin Rane. Image Segmentation Introduction Thresholding Region Splitting Region Labeling Statistical Region Description Application.
Machine Vision ENT 273 Hema C.R. Binary Image Processing Lecture 3.
Lecture(s) 3-4. Morphological Image Processing. 3/13/20162 Introduction ► ► Morphology: a branch of biology that deals with the form and structure of.
Morphological Image Processing
 After an image has been segmented into regions by methods such as those discussed in image segmentation chapter, the segmented pixels usually are represented.
(CSC 102) Lecture 30 Discrete Structures. Graphs.
Theory of Computational Complexity Probability and Computing Ryosuke Sasanuma Iwama and Ito lab M1.
Sheng-Fang Huang Chapter 11 part I.  After the image is segmented into regions, how to represent and describe these regions? ◦ In terms of its external.
De Bruijn sequences 陳柏澍 Novembers Each of the segments is one of two types, denoted by 0 and 1. Any four consecutive segments uniquely determine.
COMP 9517 Computer Vision Binary Image Analysis 4/15/2018
37TH ANNUAL AMATYC CONFERENCE A Modern Existence Proof Through Graphs
Digital Image Processing CP-7008 Lecture # 09 Morphological Image Processing Fall 2011.
CSE 554 Lecture 1: Binary Pictures
Graphs Hubert Chan (Chapter 9) [O1 Abstract Concepts]
Materi 10 Analisis Citra dan Visi Komputer
Lecture 15: Graph Theory II
R F Z D W X C E Y M S V N G T I A L H Q P J O U
CS Digital Image Processing Lecture 5
MAT 2720 Discrete Mathematics
DAGs Longin Jan Latecki
Presentation transcript:

1 Euler Number Computation By Kishore Kulkarni (Under the Guidance of Dr. Longin Jan Latecki)

2 Outline  Connected Components & Holes  Euler Number & its significance For Binary Images For Graphs Analogy between Euler number for graphs & binary images  Concavities and Convexities  Relation between Euler Number, Concavities & Convexities  Euler Number computation for test images.  Justification of correctness of algorithm used for computing Euler Number

3 Connected Components & Holes  Connected Components – 4, 6 and 8 Connectivity.  Holes

4 Euler Number The Euler Number (also called as the connectivity factor) is defined as the difference between connected components and holes. Formally, Euler Number is given by n comp E = n comp - ∑ n i hole where i = 1 n comp => number of foreground connected components n i hole => number of holes for i th connected component

5 Euler Number E = 5 – 11 = -6E = 11 – 4 = 7

6 Significance of Euler Number  Euler Number is a topological parameter, which describes the structure of an object, regardless of its specific geometric shape.  Euler number (for discrete case) is a local measure and is obtained as the sum of measures in the local neighbourhoods.  Although it is a local measure, it can be used to obtain global properties.  For objects with no holes, it can be used to calculate the number of connected components.

7 Euler Number for Graphs If V, E, and F represent the set of vertices, edges and faces respectively then for any graph we have |V| + |F| = |E|+ R where R is the Euler Number for graphs |V| = 7, |F| = 3(or 4), |E| = 9 Euler Number, R = 7 + 3(or 4) – 9 = 1(or 2) depending upon whether or not we treat the area outside the graph as a face.

8 Equivalence of Euler Number for graphs and Binary images  Any binary image can be represented as a graphs where Every black pixel (1) is represents the vertex of the graph Two neighbouring black pixels (vertices of the graphs) are joined to form an edge  The Euler Number for the binary image and its corresponding graph is the same.

9 Euler Number – Contd… Consider the following image. Euler Number E = 2 – 0 = 2. For the corresponding graph, we have |V| = 28, |E| = 30, |F| = 4 R = 28 – = 2

10 Convexities & Concavities  A region R is said to be convex iff, for every x 1, x 2 in R, [x 1, x 2 ] belongs to R.  A convex hull for a region R is defined as the convex closure of region R.  Convexity (or the convex number) N 2 + (X, α) is defined as the number of first entries in the object in a given direction α.

11 Convexity & concavity….  Concavity (or the concave number) N 2 - (X, α) is defined as the number of exits in a given direction α.  In the following figure, N 2 + (X, 90) = 3 N 2 - (X, 90) = 2

12 Convexities and Concavities…  For a binary image with 4-connectivity the convexities and concavities can be represented by the following 2*2 masks.  Convexities – Three 0’s and one 1.  Concavities – Three 1’s and one 0.

13 Euler Number in terms of convexity and concavity  The relationship between Euler Number, Convexities and concavities is given by N 2 (X) = N 2 + (X) - N 2 - (X)  So now we have the following relation E = c – h = v – e + f = N 2 + (X) - N 2 - (X)  The upstream convexity [ ] (1)  The upstream concavity [ ] (2)

14 Algorithm for computing Euler Number (for 4-connectivity) For every pixel (x, y) of the image, Do if p(x,y) = 1 and p(x-1, y) = p(x, y-1) = p(x-1, y-1) = 0 convexities ++ if p(x,y) = 0 & p(x+1, y) = p(x, y+1) = p(x+1, y+1) =1 concavities ++ end Euler number = convexities - concavities

15 Euler Number for test images Image 1: Number of Convexities = 74 Number of Concavities = 74 Euler Number = 0

16 Euler Number for test images Image 2: Number of Convexities = 86 Number of Concavities = 85 Euler Number = 1

17 Euler Number for test images Image 3: Number of Convexities = 135 Number of Concavities = ? Euler Number = ?

18 Euler Number for test images Image 4: Number of Convexities = 74 Number of Concavities = 73 Euler Number = 1

19 Euler Number for test images Image 5: Number of Convexities = 27 Number of Concavities = 16 Euler Number = 11 = Number of Objects

20 Euler Number for test images Image 6: Number of Convexities = 51 Number of Concavities = 61 Euler Number = -10 Number of Holes = 1-(-10) = 11

21 Why does this approach always returns Euler Number ? To prove: Euler Number for an image = Number of upstream convexities - Number of upstream concavities Proof: First we prove that Euler Number for the whole image is the sum of Euler numbers for every component. Then we prove the above equality for one component to complete the proof.

22 Proof - Contd…. Euler number for component is given by E i = 1 – n i hole where n i hole is the number of holes in component i Consider n comp n comp ∑ E i = ∑ (1 - n i hole ) i = 1 i = 1 n comp = n comp - ∑ n i hole i = 1 = E image

23 Proof - Contd…. The upstream convexity [ ] (1) The upstream concavity [ ] (2) The first pixel with value ‘1’ in the image should be of type (1). Hence this convexity refers to the connected component. Now any pixel with the mask [ ] can (1)be a Hole (2)be a Background If (1) then we are done..

24 Proof - Contd…. If it is a background pixel then we have an extra concavity count. Hence we need to prove that there exists an equivalent convexity to balanced the same. Now if it a background pixel, then the two edges in the 2*2 mask (referring to graphical representation of a binary image) should not form a cycle with any of the previously visited black pixels (otherwise we get a hole). If this is the case there should be a pixel with the mask [ ]. Hence we can prove that number of concavities exceeding number of holes and convexities exceeding 1 are equal by similar argument. Thus they cancel out. Hence the result.