ITERATIVE RECONSTRUCTION METHODS USING REGULARIZATION AND OPTIMAL CURRENT PATTERNS IN ELECTRICAL-IMPEDANCE TOMOGRAPHY

Electrical impedance tomography (EIT) injects currents into an object and measures the resulting voltages to determine the object's internal resistivity distribution. In medical applications, we can only make current and voltage measurements through boundary electrodes, and thus can obtain limited information. Because boundary electrodes are far from internal objects, the image reconstruction is an ill-conditioned problem. We developed an iterative reconstruction method which minimizes the effects of the ill-conditioning. Based on the modified Newton-Raphson algorithm, we developed a regularization method which integrates prior information into the image reconstruction. This improves the conditioning of the information matrix in the modified Newton-Raphson algorithm. We used optimal current patterns to obtain voltages with maximal signal-to-noise ratio (SNR). We employed a complete finite element model (FEM) for both the internal and the boundary electric fields. Reconstructed images from phantom data show that the uses of regularization, optimal current patterns, and a complete FEM model improve image accuracy. We also investigated the factors affecting the image qualilty of the iterative algorithm such as the initial guess, image iteration, and optimal current updating.