• DocumentCode
    2502849
  • Title

    Segmental reconstruction of myocardial perfusion from biplane angiographic projections by regularized iterative matrix inversion

  • Author

    Dumay, Acm ; Gerbrands, J.J. ; Zijdenbos, A.P. ; Reiber, J.H.C.

  • Author_Institution
    Lab. for Clinical & Exper. Image Processing, Lieden Univ. Hospital, Netherlands
  • fYear
    1991
  • fDate
    23-26 Sep 1991
  • Firstpage
    121
  • Lastpage
    124
  • Abstract
    The proposed segmental reconstruction technique allows the assessment of the relative mass density levels (segment values) in eight segments of slices through the myocardial muscle from biplane X-ray angiographic projections, and has been adopted from an approach by D.G.W. Onnasch et al. (1986). From a simple projection model, a set of eight dependent equations can be derived. This set is solved by regularized iterative matrix inversion that minimizes least square errors under imposed constraints. The projection images and a reconstructed slice are illustrated. The definition of the segments, intervals at projection data, and the related projection matrix are defined and the iterative regularized matrix inversion is described. Experiments with computer simulations have shown that differences between segment values of actual and reconstructed slices were less than -45% of maximum (normal) segment values
  • Keywords
    biorheology; blood; cardiology; diagnostic radiography; muscle; biplane X-ray angiographic projections; computer simulations; dependent equations; least square errors minimization; myocardial muscle slices; myocardial perfusion; projection images; regularized iterative matrix inversion; Attenuation; Electromagnetic wave absorption; Equations; Geometry; Image reconstruction; Image segmentation; Iterative methods; Least squares methods; Myocardium; X-ray imaging;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computers in Cardiology 1991, Proceedings.
  • Conference_Location
    Venice
  • Print_ISBN
    0-8186-2485-X
  • Type

    conf

  • DOI
    10.1109/CIC.1991.169060
  • Filename
    169060