1 Two dimensional thining Let P =(V, m, n, B) and P ' = (V, m, n, B - D) be digital pictures, where D B. Then we say that P' is obtained from P by deleting.

Slides:



Advertisements
Similar presentations
TWO STEP EQUATIONS 1. SOLVE FOR X 2. DO THE ADDITION STEP FIRST
Advertisements

Mathematical Preliminaries
3.6 Support Vector Machines
Advanced Piloting Cruise Plot.
Extension Principle Adriano Cruz ©2002 NCE e IM/UFRJ
1 Vorlesung Informatik 2 Algorithmen und Datenstrukturen (Parallel Algorithms) Robin Pomplun.
Chapter 1 The Study of Body Function Image PowerPoint
1 Copyright © 2013 Elsevier Inc. All rights reserved. Appendix 01.
By D. Fisher Geometric Transformations. Reflection, Rotation, or Translation 1.
Business Transaction Management Software for Application Coordination 1 Business Processes and Coordination.
Introduction to Algorithms
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
0 - 0.
DIVIDING INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
ADDING INTEGERS 1. POS. + POS. = POS. 2. NEG. + NEG. = NEG. 3. POS. + NEG. OR NEG. + POS. SUBTRACT TAKE SIGN OF BIGGER ABSOLUTE VALUE.
SUBTRACTING INTEGERS 1. CHANGE THE SUBTRACTION SIGN TO ADDITION
MULT. INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
FACTORING ax2 + bx + c Think “unfoil” Work down, Show all steps.
Addition Facts
Environmental Remote Sensing GEOG 2021
Points, Vectors, Lines, Spheres and Matrices
Introduction to digital topology
Algorithms for Geometric Covering and Piercing Problems Robert Fraser PhD defence Nov. 23, 2012.
1 © R. Guerraoui Universal constructions R. Guerraoui Distributed Programming Laboratory.
Ken C. K. Lee, Baihua Zheng, Huajing Li, Wang-Chien Lee VLDB 07 Approaching the Skyline in Z Order 1.
Minimum Weight Plastic Design For Steel-Frame Structures EN 131 Project By James Mahoney.
Randomized Algorithms Randomized Algorithms CS648 1.
Chapter 4 Systems of Linear Equations; Matrices
Parallel List Ranking Advanced Algorithms & Data Structures Lecture Theme 17 Prof. Dr. Th. Ottmann Summer Semester 2006.
ABC Technology Project
Chapter 9 -- Simplification of Sequential Circuits.
CSE554Cell ComplexesSlide 1 CSE 554 Lecture 3: Skeleton and Thinning (Part II) Fall 2013.
1 Undirected Breadth First Search F A BCG DE H 2 F A BCG DE H Queue: A get Undiscovered Fringe Finished Active 0 distance from A visit(A)
VOORBLAD.
15. Oktober Oktober Oktober 2012.
BIOLOGY AUGUST 2013 OPENING ASSIGNMENTS. AUGUST 7, 2013  Question goes here!
On Specification and Verification of Location- Based Fault Tolerant Mobile Systems Alexei Iliasov, Victor Khomenko, Maciej Koutny and Alexander Romanovsky.
© 2012 National Heart Foundation of Australia. Slide 2.
Lets play bingo!!. Calculate: MEAN Calculate: MEDIAN
Abbas Edalat Imperial College London Contains joint work with Andre Lieutier (AL) and joint work with Marko Krznaric (MK) Data Types.
Understanding Generalist Practice, 5e, Kirst-Ashman/Hull
Splines IV – B-spline Curves
6.4 Best Approximation; Least Squares
1 Cavity, concavity Cavity – bounded connected component of background (a hollow in an object) Concavity - concave shapes of the contour of an object.
Addition 1’s to 20.
25 seconds left…...
Januar MDMDFSSMDMDFSSS
Copyright © Cengage Learning. All rights reserved.
Week 1.
We will resume in: 25 Minutes.
©Brooks/Cole, 2001 Chapter 12 Derived Types-- Enumerated, Structure and Union.
PSSA Preparation.
Distributed Computing 9. Sorting - a lower bound on bit complexity Shmuel Zaks ©
9. Two Functions of Two Random Variables
Bart Jansen 1.  Problem definition  Instance: Connected graph G, positive integer k  Question: Is there a spanning tree for G with at least k leaves?
1 Decidability continued…. 2 Theorem: For a recursively enumerable language it is undecidable to determine whether is finite Proof: We will reduce the.
Equivalence Relations
Copyright © Cengage Learning. All rights reserved.
Interval Graph Test.
Secret Sharing, Matroids, and Non-Shannon Information Inequalities.
A. S. Morse Yale University University of Minnesota June 4, 2014 TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A.
1 © R. Guerraoui The Limitations of Registers R. Guerraoui Distributed Programming Laboratory.
Introduction Image geometry studies rotation, translation, scaling, distortion, etc. Image topology studies, e.g., (i) the number of occurrences.
Relationship between pixels Neighbors of a pixel – 4-neighbors (N,S,W,E pixels) == N 4 (p). A pixel p at coordinates (x,y) has four horizontal and vertical.
Digital Image Processing Lecture 13: Image Topology - Skeletonization
Presentation transcript:

1 Two dimensional thining Let P =(V, m, n, B) and P ' = (V, m, n, B - D) be digital pictures, where D B. Then we say that P' is obtained from P by deleting the points in D. Alternatively, we may say that P is obtained from P' by adding the points in D. Image thinning is a common pre-processing operation in pattern recognition. Its goal is to reduce the set of black points to a "skeleton" in a "topology- preserving" way. Effect of a thinning algorithm on the (8, 4) digital picture

2 Maximal ball Let denote by B r (x) the ball of radius r (strictly positive integer) centred on x є Z 2, defined by B r (x) = {y є Z 2, d(x, y) r}, where function d:Z 2 R + {0} is a metric. Let assume a digital image (Z 2, m, n, B). A ball B r (x) B is maximal for B if it is not strictly included in any other ball included in B. The medial axis of B, denoted by MA(B), is the set of the centres of all the maximal balls for B. Examples of balls of radiuses 1, 2, 3 respectively (city distance)

3 Example of medial axis [Malina 02] p. 83

4 Medial axis vs thining and shrinking A non-topological requirement of a thinning algorithm is that each elongated part of the input black point set should be represented by a black digital arc in the output skeleton. An algorithm which does not meet this condition, but merely deletes black points while preserving the topology of the image, is called a shrinking algorithm. Input objectShrinking resultMedial axis Thining result One iteration of thining [Coup 07]

5 Topology preservation criterion Let P = (Z 2, m, n, B) P' = (Z 2, m, n, B - D) be a two-dimensional digital pictures. Then deletion of the points in a subset D of B preserves topology if and only if each black component of P contains exactly one black component of P', and each white component of P' contains exactly one white component of P, PP P P Example 1: (8,4) dig. pic.Example 2: (8,4) dig. pic.

6 Topology preservation criterion Let P = (Z 2, m, n, B) be a two-dimensional digital picture. Then deletion of the points in a subset D of B preserves topology if and only if each black component of P contains exactly one black component of P', and each white component of P' contains exactly one white component of P, where P' is the digital picture (Z 2, m, n, B - D). PP Example 3: (8,4) dig. pic.

7 Simple point A black point p in a two-dimensional digital picture is called a simple point if its deletion preserves topology in the sense of Criterion from the previous slide. a b c [Coup 10]

8 Theorem on simple points in 2D Let p be a non-isolated border point in an (8, 4) or (4, 8) digital picture. Let B be the black point set of the digital picture and let B' = B { p}. Then p is a simple point iff: p is adjacent to just one component of N 8 (p) B'. p is adjacent to just one component of N 8 (p) B *, where B * = Z 2 \ B Where N 8 (p) is a 8-neighbourhood of p. (def. in slide 7). Proof in () a Let introduce: T(p) – number of components of N 8 (p) B'. Tb(p) – number of components of N 8 (p) B* So: T(a) = Tb(a) = 1; T( ) = Tb( ) = 2; T( )=0, Tb( )=1 Let p be a non-isolated border point in an (8, 4) or (4, 8) digital picture. Then p is a simple point iff T(p) = 1 and Tb(p) = 1

9 Sequential deletion of simple points Let P 0, P 1,..., P n be a sequence of digital pictures. If for each 1 < i < n the picture P i+1 is obtained by deleting a simple point of P i, from P i then we say that P n is obtainable from P 0 (or that P 0 can be transformed into P n ) by sequential deletion of simple points. Important! A simple point of P i, need not be a simple point of P 0 ; A black point of P i that is a simple point of P 0 need not be a simple point of P i

10 Example of sequential deletion of simple points P0P0 P1P1 P2P2 P3P3 P4P4 P5P5 P6P6 P7P7 P8P8 P9P9

11 Problem with parallel deletion of simple points p q P0P0 P1P1 P2P2 P3P3 P4P4 P5P5 P6P6 P7P7 P8P8 P9P9 s t

12 Theorems about sequential deletion of simple points Any finite (4, 8) or (8, 4) digital picture whose black point set is non-empty and connected and has no holes can be transformed by sequential deletion of simple points to a digital picture with just one black point. [Rose 70] Finite (4, 8) digital picture whose black point set is connected and has just one hole can be transformed by sequential deletion of simple points to a digital picture whose black point set is a simple closed black curve. [Rose 73] Finite (4, 8) and (8, 4) digital picture topology preservation in the sense of the criterion from slide 28 is equivalent to the condition that P' be obtainable from P by sequential deletion of simple points. [Rose 98] In any finite (4, 8) digital picture sequential deletion of simple points will eventually produce a digital picture whose black point set does not contain any 2 by 3 arrays of black points. [Alex 71] Given two finite (4, 8) digital pictures whose black point sets are connected, and which have the same number of holes, it is possible to transform one to the other by sequential addition and deletion of simple points. [Mylo 71]

13 2D parallel thinning algorithms It is generally quite tricky to prove that a proposed parallel thinning algorithm satisfies the criterion from slide 28. For an example of such a proof see [Stef 71]. Unfortunately such proofs often have to be done from first principles.

14 Theorem on border parallel deletion of points A black point with coordinates (x, y) is said to be a north border point if the point (*, y + 1) is a white point. An end point of a two-dimensional digital picture is a black point that is adjacent to just one other black point. Theorem 2. Let P be an (8, 4) or a (4, 8) digital picture. Then (parallel) deletion of any number of simple north border non-end points of P preserves topology in the sense of criterion from slide 27. Simple north-border non-end points in an (8, 4) digital picture

15 Border sequential Theorem 2 obviously remains valid if "east," "south," or "west" is substituted for "north." However, the restriction that border points are deleted from just one side (north) is necessary. p q Simple non-end points in an (8, 4) digital picture whose parallel deletion will merge two white components Theorem 2 is applicable to algorithms which delete points in parallel from each side in turn (e.g., in the order N, S, E, W). Such algorithms have been called border sequential [Hild 83].

16 Border parallel algorithms Theorem 2 () can also be applied to so-called border parallel algorithms [Hild 83]. These algorithms are built up of passes each consisting of four subcycles: in each subcycle points are removed from just one side, but a point p is deleted only if the point deletion criterion is satisfied both by ps current 3 by 3 neighborhood N(p) and by N(p) as it was at the start of the current pass. Dobrze byłoby dać przykład. Jeżeli się nie da to raczej usunąć. Problem z tym, że nie mogę znaleźć artykułu [44] autorstwa Hilditch J C

17 Topological characteristic of points in 2D A latice point p is: Simple point: T(p) = 1 Tb(p) = 1 (see slide 30)see slide 30 Interior point: Tb(p) = 0 Isolated point T(p) = 0 Curve point T(p) = 2 and Tb(p) = 2 Curve junction point T(p)=3 or Tb(p)=3 Interior point border point curve point junction point between curves Isolated point

18 Theorem about simple points in 3D Let X Z 3 and x X. K 6 (x, X) –number of components of N 18 (x) X \ {x} adjacent to x. K 26 (x, X) –number of components of N 26 (p) X \{x} adjacent to x. Let p be u non-isolated border point in an (m, n) digital picture. Let B be the black point set of the digital picture. Then p is a simple point iff K m (p, B) = 1 and K n (p, B * ) = 1, where B * = Z 3 - B

19 Topological numbers T, Tb in 3D In 3D we similarly calculate T(p) and Tb(p), for a black point p and (Z3, 26, 6, B) digital image. T(p) = K 26 (p, B) and Tb(p) = K 6 (p, B * ) T(p) is calculated as a number of black 26-components in N 26 (p) - {p} N 26 (p)-{p} p T(p) = 1

20 Topological numbers T, Tb in 3D Tb(p) is calculated as a number of white 6-components in N 18 (p) 6-adjacent to p. p N 18 (p) Tb(p) = 3

21 Topological characteristic of points in 3D For a 3D (26, 6) digital image a point p is: An interior point Tb(p) = 0 A isolated pointT(p) = 0 A border point T(p) = 1 and Tb(p) = 1 A curve pointT(p) = 2 and Tb(p) = 1 A curves junctionT(p) > 2 and Tb(p) = 1 A surface pointT(p) = 1 and Tb(p) = 2 A surface-curve(s) junctionT(p) 2and Tb(p) = 1 A surfaces junctionT(p) = 1 and Tb(p) > 2 A surfaces-curve(s) junctionT(p) 2and Tb(p) > 2 Junction between curves curve Border points Junction between surfaces surface Junction curve-surface [Malan 10]

22 Example Fragment of (26, 6) image. Junction between surfaces: T(p) = 1 and Tb(p) > 2 Calculation of T(p) p N 26 (p)-{p} T(p) = 1 p N 18 (p) Tb(p) = 4 Calculation of Tb(p)