Numerical implementation of a variational method for electrical impedance tomography

Robert Kohn, A. McKenney

Research output: Contribution to journalArticle

Abstract

A variational method for computing electrical conductivity distributions from boundary measurements was proposed by Kohn and Vogelius (1987). The authors explore the numerical performance of that technique. A version of Newton's method is used for the minimisation, and synthetic data for the boundary measurements. The variational method is found to be generally stable and robust, reproducing the locations and shapes of conducting objects well, provided that smooth boundary data are used. Early termination appears to have a desirable smoothing effect upon the reconstruction. Contrary to the suggestion of Kohn and Vogelius, the method is not enhanced by allowing the conductivity to be anisotropic.

Original languageEnglish (US)
Article number009
Pages (from-to)389-414
Number of pages26
JournalInverse Problems
Volume6
Issue number3
DOIs
StatePublished - 1990

Fingerprint

Electrical Impedance Tomography
Acoustic impedance
electrical impedance
Variational Methods
Tomography
tomography
Newton-Raphson method
Early Termination
Smoothing Effect
Newton methods
Electrical Conductivity
Synthetic Data
smoothing
Newton Methods
Conductivity
suggestion
conduction
conductivity
electrical resistivity
optimization

ASJC Scopus subject areas

  • Applied Mathematics
  • Mathematical Physics
  • Physics and Astronomy(all)
  • Statistical and Nonlinear Physics

Cite this

Numerical implementation of a variational method for electrical impedance tomography. / Kohn, Robert; McKenney, A.

In: Inverse Problems, Vol. 6, No. 3, 009, 1990, p. 389-414.

Research output: Contribution to journalArticle

@article{cb8d40ee9155444092f6182cde4102c7,
title = "Numerical implementation of a variational method for electrical impedance tomography",
abstract = "A variational method for computing electrical conductivity distributions from boundary measurements was proposed by Kohn and Vogelius (1987). The authors explore the numerical performance of that technique. A version of Newton's method is used for the minimisation, and synthetic data for the boundary measurements. The variational method is found to be generally stable and robust, reproducing the locations and shapes of conducting objects well, provided that smooth boundary data are used. Early termination appears to have a desirable smoothing effect upon the reconstruction. Contrary to the suggestion of Kohn and Vogelius, the method is not enhanced by allowing the conductivity to be anisotropic.",
author = "Robert Kohn and A. McKenney",
year = "1990",
doi = "10.1088/0266-5611/6/3/009",
language = "English (US)",
volume = "6",
pages = "389--414",
journal = "Inverse Problems",
issn = "0266-5611",
publisher = "IOP Publishing Ltd.",
number = "3",

}

TY - JOUR

T1 - Numerical implementation of a variational method for electrical impedance tomography

AU - Kohn, Robert

AU - McKenney, A.

PY - 1990

Y1 - 1990

N2 - A variational method for computing electrical conductivity distributions from boundary measurements was proposed by Kohn and Vogelius (1987). The authors explore the numerical performance of that technique. A version of Newton's method is used for the minimisation, and synthetic data for the boundary measurements. The variational method is found to be generally stable and robust, reproducing the locations and shapes of conducting objects well, provided that smooth boundary data are used. Early termination appears to have a desirable smoothing effect upon the reconstruction. Contrary to the suggestion of Kohn and Vogelius, the method is not enhanced by allowing the conductivity to be anisotropic.

AB - A variational method for computing electrical conductivity distributions from boundary measurements was proposed by Kohn and Vogelius (1987). The authors explore the numerical performance of that technique. A version of Newton's method is used for the minimisation, and synthetic data for the boundary measurements. The variational method is found to be generally stable and robust, reproducing the locations and shapes of conducting objects well, provided that smooth boundary data are used. Early termination appears to have a desirable smoothing effect upon the reconstruction. Contrary to the suggestion of Kohn and Vogelius, the method is not enhanced by allowing the conductivity to be anisotropic.

UR - http://www.scopus.com/inward/record.url?scp=0000430260&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0000430260&partnerID=8YFLogxK

U2 - 10.1088/0266-5611/6/3/009

DO - 10.1088/0266-5611/6/3/009

M3 - Article

VL - 6

SP - 389

EP - 414

JO - Inverse Problems

JF - Inverse Problems

SN - 0266-5611

IS - 3

M1 - 009

ER -