Displaced Electrodes in EIT and Their Influence on Reconstructed Absolute Images

Abstract

In the process of obtaining absolute images through Electrical Impedance Tomography (EIT), it is necessary to deal with various sources of error. The reconstruction of absolute images in EIT is only feasible with the use of prior information. In the reconstruction of these images, it is assumed that the position of the electrodes is known. However, this assumption does not always correspond to reality, as the electrodes may be displaced from their ideal positions. This project investigates the feasibility of correcting the electrical potentials in the presence of displaced electrodes, using prior information. For this purpose, both supervised and unsupervised neural networks will be explored. Three approaches to solve the problem of displaced electrodes will be examined: 1. Identification and correction of the displaced electrode; 2. Generic correction of displaced electrodes; and 3. Image reconstruction without making corrections for electrode displacement. (AU)