Defuzzification using Steiner points


Vetterlein, Thomas, Navara, Mirko
Fuzzy Sets and Systems
Volume 157, Number 11, Pages 1455-1462
June, 2006

Abstract

A defuzzification function assigns to each fuzzy set a crisp value in a way that this value may intuitively be understood as the ``centre'' of the fuzzy set. In the present paper, this vague concept is put into a mathematically rigorous form. To this end, we proceed analogously to the case of sharply bordered subsets, for which the Steiner point is frequently used. The function assigning to each convex subset its Steiner point is characterised by three properties; here, we study functions whose domains consist of fuzzy sets and which fulfil analogous properties. Although uniqueness can no longer be achieved, we give a complete characterisation of what we call Steiner points of fuzzy sets.

Keywords

fuzzy set, defuzzification, support function, Steiner point, computer vision, medical imaging

Full Paper

Portable document format file PDF
Gzipped postscript file ps.gz ()


Bibtex entry

@Article{VetterleinNavara-FSS_Steiner,
  author =     {Vetterlein, Thomas and Navara, Mirko},
  title =      {Defuzzification using {S}teiner points},
  year =       {2006},
  month =      {June},
  pages =      {1455--1462},
  journal =    {Fuzzy Sets and Systems},
  publisher =  { Elsevier Science },
  address =    { Amsterdam, The Netherlands },
  issn =       { 0165-0114 },
  authorship = { 50-50 },
  volume =     {157},
  number =     {11},
  annote = {A defuzzification function assigns to each fuzzy set a
    crisp value in a way that this value may intuitively be understood
    as the ``centre'' of the fuzzy set.  In the present paper, this
    vague concept is put into a mathematically rigorous form. To this
    end, we proceed analogously to the case of sharply bordered
    subsets, for which the Steiner point is frequently used. The
    function assigning to each convex subset its Steiner point is
    characterised by three properties; here, we study functions whose
    domains consist of fuzzy sets and which fulfil analogous
    properties.  Although uniqueness can no longer be achieved, we
    give a complete characterisation of what we call Steiner points of
    fuzzy sets.},
  keywords =   {fuzzy set, defuzzification, support function, 
    Steiner point, computer vision, medical imaging},
  psurl = { [PDF] },
}