Cole Parameter Estimation for Calibrating the Electrical Impedance Tomography System

Abstract

Impedance measurement inaccuracies and difficulties in modeling the complex electrical properties of biological tissues often compromise the diagnostic reliability of Electrical Impedance Tomography (EIT). To address these limitations, this study proposes a comprehensive calibration procedure that enhances EIT image quality through the accurate estimation of Cole model parameters. The procedure first characterizes the frequency-dependent non-idealities of the measurement hardware. An optimization algorithm then estimates the Cole parameters of a reference phantom, generating a system-wide correction map that is applied to subsequent measurements. Four optimization algorithms, Levenberg-Marquardt (LM), Nelder-Mead Simplex, Interior-Point, and Active-Set, were evaluated for noise robustness, computational efficiency, and constraint handling. We validated the procedure's efficacy through simulations, in which the LM algorithm demonstrated superior performance with a root-mean-square error of 0.40 Omega at 20% noise. Experimental measurements on a resistor phantom network further confirmed this approach, yielding an average impedance estimation error of 2.4%. Validation with a lung phantom resulted in substantially improved imaging quality, marked by enhanced anomaly detection and reduced artifacts post-calibration. These findings demonstrate that the proposed calibration procedure effectively improves the reliability of EIT systems for critical biomedical applications, including lung monitoring and tumor detection.