• DocumentCode
    3063185
  • Title

    A new inconsistency measure for linear systems and two applications in motion analysis

  • Author

    Niu, Yan ; Dick, Anthony ; Brooks, Michael

  • Author_Institution
    Sch. of Comput. Sci., Univ. of Adelaide, Adelaide, SA, Australia
  • fYear
    2009
  • fDate
    23-25 Nov. 2009
  • Firstpage
    12
  • Lastpage
    17
  • Abstract
    This paper proposes a new measure that continuously measures the degree of inconsistency for a linear system. We apply the new measure to two essential vision problems. One is to predict the fidelity of a local optical flow computation system; The other is to detect motion boundaries. Experimental results on benchmark sequences validate the performance of the proposed measure on both problems.
  • Keywords
    image sequences; linear systems; motion estimation; inconsistency measure; linear systems inconsistency degree; motion analysis; motion boundary detection; Australia; Computer science; Computer vision; Fluid flow measurement; Image motion analysis; Linear systems; Motion analysis; Motion measurement; Optical computing; Vectors; Optical flow; confidence measure; inconsistent linear system; motion boundary;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Image and Vision Computing New Zealand, 2009. IVCNZ '09. 24th International Conference
  • Conference_Location
    Wellington
  • ISSN
    2151-2205
  • Print_ISBN
    978-1-4244-4697-1
  • Electronic_ISBN
    2151-2205
  • Type

    conf

  • DOI
    10.1109/IVCNZ.2009.5378427
  • Filename
    5378427