• DocumentCode
    1119304
  • Title

    A Method for Computing the Partial Singular Value Decomposition

  • Author

    Shlien, Seymour

  • Author_Institution
    Department of Communications, Communications Research Center, Box 11490, Station H, Ottawa, Ont., Canada K2H 8S2.
  • Issue
    6
  • fYear
    1982
  • Firstpage
    671
  • Lastpage
    676
  • Abstract
    A method for computing the partial singular value decomposition of a matrix is described. The method is appropriate to problems where the matrix is known to be of low rank and only the principal singular vectors are of interest. The technique is simple, easy to implement in integer arithmetic, and places modest memory requirements. The convergence properties of the algorithm are investigated analytically and by simulation.
  • Keywords
    Arithmetic; Biology computing; Geophysics computing; Image processing; Marine animals; Matrices; Matrix decomposition; Pattern recognition; Signal processing; Singular value decomposition; Convergence properties; SVD; eigenvalue/eigenvector; image processing; numerical methods; power method;
  • fLanguage
    English
  • Journal_Title
    Pattern Analysis and Machine Intelligence, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0162-8828
  • Type

    jour

  • DOI
    10.1109/TPAMI.1982.4767324
  • Filename
    4767324
    • Link To Document
    https://search.isc.ac/dl/search/defaultta.aspx?DTC=49&DC=1119304