• DocumentCode
    374854
  • Title

    A simultaneous reconstruction/segmentation method for 123I SPECT striata imaging

  • Author

    Laurette, I. ; Debreuve, E. ; Darcourt, J. ; Barland, M.

  • Author_Institution
    Lab. de Biophys. et Traitement de l´´Image, Nice Univ., Sophia-Antipolis, France
  • Volume
    2
  • fYear
    2000
  • fDate
    2000
  • Abstract
    For diagnosis purpose, quantification of 123I labeled dopaminergic tracers in striata regions is required. The quantification is often performed by manually or automatically choosing regions of interest in striata to calculate absolute or relative bindings. This segmentation step is inextricably linked to the quality of reconstruction process. It is well known that SPECT suffers from poor data quality, imposing the use of regularization such as low-pass filtering. Thus, the detection of organ edges is made difficult. The authors propose a simultaneous reconstruction/segmentation method which iteratively warps some initial shapes to the desired objects. The method was tested on an anthropomorphic striatal phantom, for the detection of the striatal compartments
  • Keywords
    brain; image reconstruction; image segmentation; medical image processing; single photon emission computed tomography; 123I SPECT striata imaging; 123I labeled dopaminergic tracers quantification; I; anthropomorphic striatal phantom; low-pass filtering; medical diagnostic imaging; nuclear medicine; organ edges detection; poor data quality; reconstruction process quality; simultaneous reconstruction/segmentation method; striatal compartments detection; Image edge detection; Image reconstruction; Image segmentation; Imaging phantoms; Level set; Low pass filters; Performance analysis; Shape; Testing; Topology;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Nuclear Science Symposium Conference Record, 2000 IEEE
  • Conference_Location
    Lyon
  • ISSN
    1082-3654
  • Print_ISBN
    0-7803-6503-8
  • Type

    conf

  • DOI
    10.1109/NSSMIC.2000.950098
  • Filename
    950098