Analysing the performance of EIT images using the point spread function

Abstract

Electrical impedance tomography (EIT) is a noninvasive medical imaging technique in which a small current is applied to the electrodes attached to the surface of a subject body and a cross-sectional image of the resistivity (or conductivity) distribution inside the body is reconstructed using an inverse algorithm. Due to the ill-posed nature of EIT inverse problem, EIT bears poor spatial resolution and behaves non-linearly in nature. Point spread function (PSF), which is calculated over the whole domain as responses to a small circular anomaly moving around the entire domain, is a characteristic parameter to estimate the performance of imaging systems. In order to analyze the quality of EIT reconstructed image, PSF is employed in this work. PSF incorporates the key imaging attributes, comprising amplitude response, resolution, position error, shape deformation and ringing effect. This paper presents a numerical study on the use of PSF for static and dynamic EIT image reconstruction. The static image reconstruction is done using the modified Newton Raphson (mNR) algorithm whereas the dynamic image reconstruction is done with extended Kalman filter (EKF). A detailed analysis of the performance of mNR and EKF has been carried out based upon on the imaging attributes gathered using the said algorithms. The results are convincing and provide a fresh perspective to use the PSF in order to analyze the performance of EIT image reconstruction.