• DocumentCode
    3086705
  • Title

    The advantages of multiprocessor systems for ACEIT and ICEIT inverse problem solution

  • Author

    Kacarska, Marija ; Loskovska, S. ; Andonov, D.

  • Author_Institution
    Fac. of Electr. Eng., Univ. Sv. Kiril i Metodij, Skopje, Macedonia
  • Volume
    5
  • fYear
    2001
  • fDate
    2001
  • Firstpage
    279
  • Abstract
    In this paper two multiprocessor systems for ACEIT and ICEIT inverse problem solution are presented. Pipeline Multiprocessor Algorithm (PMA) and an algorithm based on systolic arrays are compared with standard EIT reconstruction algorithms. Discussions consider time and resource requirements for image reconstruction and spatial resolutions for each of proposed algorithms. Multiprocessor based algorithms solve the problem using the full set of parameters, providing good spatial resolution of the images, and calculate the solution faster. According to the efficiency, PMA has the advantage over systolic arrays, because the number of processors in the system is very small and the processor load balance is optimized
  • Keywords
    computerised tomography; electric impedance imaging; image reconstruction; inverse problems; medical image processing; multiprocessing systems; pipeline processing; systolic arrays; ACEIT; ICEIT; applied-current electrical impedance tomography; image reconstruction; induced-current electrical impedance tomography; inverse problem solution; multiprocessor systems; pipeline multiprocessor algorithm; processor load balance; spatial resolutions; systolic arrays; Data acquisition; Electrodes; Inverse problems; Magnetic field measurement; Multiprocessing systems; Reconstruction algorithms; Spatial resolution; Systolic arrays; Tomography; Voltage;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Circuits and Systems, 2001. ISCAS 2001. The 2001 IEEE International Symposium on
  • Conference_Location
    Sydney, NSW
  • Print_ISBN
    0-7803-6685-9
  • Type

    conf

  • DOI
    10.1109/ISCAS.2001.922039
  • Filename
    922039