Reconstruction algorithms in Medical Electrical Impedance Tomography (EIT)

Abstract

applications, such as thorax, brain, and breast screening. The principle of it is to reconstruct an image of the region of interest by mapping the conductivity distribution of the area. Among all the medical imaging technologies, EIT has the advantages of its low cost, mobility and non-invasive nature. However, it suffers from being very low resolution. and this project aims to investigate the feasibility of improvement of EIT application in breast cancer detecting.

Since EIT requires direct contact with the boundary of the region of interest, electrode positioning and boundary movement have always been considered as two of the potential sources of errors and artefacts. With the aim of turning the issues with the boundary into opportunities, a hypothesis is introduced for the first time. A breast can be deformed due to its natural structure. Therefore, if the breast is deformed intentionally, each deformation can provide a new set of independent EIT measurement data. More independent data would provide more information from the same region of interest. In this hypothesis, information gathered with different deformations is combined; in all cases, we assumed that the shape and electrode positions were measured by other means.

Simulation studies have been carried out to analyse the potential of the deformable EIT concept. Initial results show improvements in the detectability of the early stage tumour in depth. Physical experiments have been carried out to demonstrate and validate the results of the simulation. Results are encouraging, making them worthwhile for further work.

Keywords: Electrical impedance tomography, breast cancer imaging, deformable boundary

Date of Award	19 Nov 2019
Original language	English
Awarding Institution	University of Bath
Supervisor	Manuchehr Soleimani (Supervisor) & Ivan Astin (Supervisor)