002. Separation of respiration and perfusion in Electrical Impedance Tomography (EIT) images using Gaussian process regression

Electrical impedance tomography (EIT) is a non-invasive type of medical imaging in which the electrical  impedance of a part of the body is inferred from surface electrode measurements and used to form image series of that part. In this proposal we will investigate the separation of heart and lung contributions in EIT measurements of a patient’s chest. The image series can be decomposed into one-dimensional time series on which signal processing can be applied. The idea is quite simple: the respiration and heartbeat frequencies are different, hence one can distinguish the signals in the frequency domain. Modelling the signals will be done by means of Gaussian processes, and hence, making the technique fully automatic. Afterwards one can reconstruct the image series for respiration and perfusion separately (see figure). The ultimate goal is online monitoring of the state of health of the lungs of a patient kept under artificial coma.

Standard deviation
Prerequisites: good knowledge in signal processing, measurement, data-based modelling and Matlab.

2021
Alberto Battistel, Hochschule Furtwangen University (HFU), Germany
Back to top