Solution of the inverse problem of magnetic induction tomography (MIT) with multiple objects: Analysis of detectability and statistical properties with respect to the reconstructed conducting region

Robert Merwa, Patricia Brunner, Andreas Missner, Karl Hollaus, Hermann Scharfetter

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)

Abstract

Magnetic induction tomography (MIT) is a technique to image the passive electrical properties (i.e. conductivity, permittivity, permeability) of biological tissues. The inverse eddy current problem is nonlinear and ill-posed, thus a Gauss-Newton one-step method in combination with four different regularization schemes is used to obtain stable solutions. Simulations with 16 excitation coils, 32 receiving coils and different spherical perturbations inside a homogeneous cylinder were computed. In order to compare the statistical properties of the reconstructed results a Monte Carlo study with a SNR of 40 dB and 20 dB was carried out. Simulated conductivity perturbations inside a homogeneous cylinder can be localized and resolved and the results prove the feasibility of difference imaging with MIT.

Original languageEnglish
Article numberS21
Pages (from-to)S249-S259
JournalPhysiological Measurement
Volume27
Issue number5
DOIs
Publication statusPublished - 1 May 2006

Keywords

  • Inverse problem
  • Magnetic induction tomography
  • Regularization
  • Reproducibility of Results
  • Electric Impedance
  • Models, Statistical
  • Magnetics
  • Feasibility Studies
  • Plethysmography, Impedance/methods
  • Algorithms
  • Models, Biological
  • Computer Simulation
  • Sensitivity and Specificity
  • Image Interpretation, Computer-Assisted/methods
  • Tomography/methods
  • Monte Carlo Method

Fingerprint Dive into the research topics of 'Solution of the inverse problem of magnetic induction tomography (MIT) with multiple objects: Analysis of detectability and statistical properties with respect to the reconstructed conducting region'. Together they form a unique fingerprint.

Cite this