Web-Books
in the Austria-Forum
Austria-Forum
Web-Books
Informatik
Joint Austrian Computer Vision and Robotics Workshop 2020
Page - 138 -
  • User
  • Version
    • full version
    • text only version
  • Language
    • Deutsch - German
    • English

Page - 138 - in Joint Austrian Computer Vision and Robotics Workshop 2020

Image of the Page - 138 -

Image of the Page - 138 - in Joint Austrian Computer Vision and Robotics Workshop 2020

Text of the Page - 138 -

clidean topology and f will describe a continuous image or3D-scan. Definition2.1. Apath isacontinuousfunctionfrom the real interval [a,b] (witha<b) into a topological spaceΩ. Definition 2.2. Two points x 6= y in a topological spaceΩarecalledpath-connected ifandonlyif there exists a path γ : [a,b]→ Ω with γ(a) = x and γ(b) =y. Definition 2.3. The set of all points which are path- connected to a pointxāˆˆā„¦ is the connected compo- nentofx: [x] :={yāˆˆā„¦|x ispath-connected toy} Any subset of Ω which can be written in above way (forasuitablechoiceofx) iscalledaconnectedcom- ponent. Definition2.4. Apathγ : [a,b]→Ω iscalledmono- tonic if andonly if thewholepath isascendingor the wholepathisdescending,meaningthefirstorsecond formula belowhas tohold, respectively: āˆ€s,t∈ [a,b] :s<t⇒f(γ(s))≤f(γ(t)) āˆ€s,t∈ [a,b] :s<t⇒f(γ(s))≄f(γ(t)) Definition 2.5. LetR āŠ‚ Ω. R is called slope re- gionormonotonicallyconnected if andonly if forall x,y∈R there exists a monotonic pathγ : [a,b]→ Rwithγ(a) =xandγ(b) =y. Definition 2.6. A family of sets{AiāŠ‚ā„¦ | i∈ I} is called a sloperegiondecomposition if andonly if: • Ai is a slope region for all i∈ I • āˆ€i,j∈ I : i 6= j⇒Ai∩Aj=āˆ…. • ā‹ƒi∈IAi= Ω Definition 2.7. Consider two slope region de- compositions A = {AiāŠ‚ā„¦ | i∈ I} and B = {BjāŠ‚ā„¦ | j∈J}. We callA coarser thanB, al- ternativelyBfiner thanA, in SymbolsA B if and only if āˆ€j∈J ∃i∈ I :BjāŠ‚Ai. Theorem2.8. isapartialorder, i.e. fulfills reflex- ivity, antisymmetryand transitivity. Proof: Straight forward. Antisymmetry follows fromthedecompositionproperty. Definition 2.9. A slope region decompositionA is calledmaximallycoarseorsimplycoarse ifandonly if there is no other coarser slope region decomposi- tion. We can apply Zorn’s lemma [6] to the partial or- der , which yields the existence of maximal ele- ments. For this we need to show that chains have upperbounds. Theorem2.10. Foranyascendingchainof slopere- gion decompositions (Ai)i∈I, that is t≄ s⇒At As, there is a slope region decompositionAāˆž satis- fyingāˆ€i∈ I :Aāˆž Ai. Proof: Weconsider theequivalencerelationā€con- nected inAiā€ for twopointsx,yāˆˆā„¦: x∼i y :ā‡”āˆƒA∈Ai :x∈A∧y∈A Theequivalence relation is a subsetofΩ2, andAt As implies∼t⊃∼s. This suggests the use ofāˆ¼āˆž:=ā‹ƒ i∈I ∼i to get an upper bound. Indeed the equiv- alence classes ofāˆ¼āˆž yield a partitionAāˆž of Ω, which is coarser than anyAi. But do they form a slope region decomposition? Yes: For any two fixed pointsx,y to beāˆ¼āˆž-connected, they need to be∼i- connected for some i ∈ I. So there is a mono- tonic path linkingx and y inA= [x]∼i āŠ‚ [x]āˆ¼āˆž, by which they are monotonically connected inAāˆž. ThereforAāˆž is a slope regiondecomposition. HenceeverysetΩhasacoarsedecomposition. Theorem 2.11. LetAāŠ‚ā„¦ be a path-connected set. A is a slope region if and only if all levelsets off in Aarepath-connected, i.e. āˆ€c∈R :fāˆ’1({c})∩A ispath-connected. Proof: ā€ā‡’ā€viacontraposition: Supposethereexistsac∈RwithL :=fāˆ’1(c)∩A not path-connected. We decomposeL in its compo- nents and pick x and y from different components. Since f(x) = f(y) = c a monotonic path between xandywouldhave to lie completely inL.However, since x and y are from different components, they cannotbeconnectedbyapathinLandthereforecan- not be connected with a monotonic path. Therefore, A isnot a slope region. ā€ā‡ā€via ironingout anarbitrarypath: Givenx,y∈Awe have to find a monotonic path γ. Without loss of generality supposef(x)≄ f(y). SinceA is path-connected, there exists an (not nec- essarily monotonic) path γ0 : [a,b]→A fromx to 138
back to the  book Joint Austrian Computer Vision and Robotics Workshop 2020"
Joint Austrian Computer Vision and Robotics Workshop 2020
Title
Joint Austrian Computer Vision and Robotics Workshop 2020
Editor
Graz University of Technology
Location
Graz
Date
2020
Language
English
License
CC BY 4.0
ISBN
978-3-85125-752-6
Size
21.0 x 29.7 cm
Pages
188
Categories
Informatik
Technik
Web-Books
Library
Privacy
Imprint
Austria-Forum
Austria-Forum
Web-Books
Joint Austrian Computer Vision and Robotics Workshop 2020