Edge-promoting priors in electrical impedance tomography
Public examination of a doctoral dissertation in the field of Applied Physics
Doctoral candidate MSc Gerardo Del Muro González
Date and venue: 8.12.2017 at 12 noon, SN200, Snellmania, Kuopio campus
Language of the dissertation and the public examination: English
In electrical impedance tomography (EIT), the conductivity and permittivity distributions (admittivity) of a target are reconstructed based on voltage measurements, known current injections, and knowledge of the target geometry. Due to the ill-posedness of this reconstruction problem, the determination of a meaningful solution depends heavily on the prior information related to the target.
In the Bayesian framework, the EIT inverse problem is formulated as a statistical inference problem. This problem is based on the statistical considerations of the a priori information about the admittivity and the noise statistics of the measurements. In this approach, the prior information can be, for example, sharp internal boundaries in the admittivity distribution, caused by the interfaces between different materials.
In his PhD thesis work, Gerardo González, MSc, has investigated prior models based on the total variation (TV) functional to improve the quality of conductivity and permittivity reconstructions. The findings presented in this thesis demonstrate the feasibility of the proposed prior models for removing high oscillations in the admittivity distribution, whilst preserving internal boundaries in such reconstructions.
The doctoral dissertation of MSc Gerardo Del Muro González, entitled Edge-promoting priors in electrical impedance tomography will be examined at the Faculty of Science and Forestry. The opponent in the public examination will be Professor Daniel Watzenig, Institute of Electrical Measurement, Graz University of Technology, Austria and the custos will be Professor Marko Vauhkonen, University of Eastern Finland.
Photo available for download at https://kuvapankki.uef.fi/A/UEF+kuvahakemisto/12563?encoding=UTF-8