Research On The Key Problems Of Magnetic Detection Electrical Impedance Tomography (MDEIT)

The body can be considered as a composite volume conductor with comprisetissues with differing electrical properties. The conductivity of biological tissueschanges for different tissues, physiological and pathological states about anatomy,physiological processes and pathology to examine human body. The electricalimpedance tomography (EIT) is an imaging technique to reconstruct the conductivitydistributions of the human body. However, due to the insufficient measured boundaryvoltage information, the image reconstruction problem of EIT is ill-posed and it canonly yield low resolution conductivity images.Magnetic detection electrical impedance tomography (MDEIT) is technicallybased on injecting an alternating current into the object with surface electrodes. Themagnetic flux density measurements recorded with magnetic field sensors around theobject are used to reconstruct the conductivity images. Since measurement of themagnetic flux density does not require surface contact, MDEIT have the precisedetector positions and can record a greater number of measurements. MDEITovercomes the shortcomings of EIT and retains its merits. The paper introduces theforward problem, the inverse problem, the optimization of measurement arrangement,and the data acquisition system in MDEIT, which establish the foundation for MDEITdevelopment.This paper describes the forward problem in MDEIT. The forward solvercomputes voltage and current density both within the object, and magnetic fluxdensity distribution surrounding the object. Based on the forward solver, differentelectrode modes are compared and an annular electrode mode is proposed to beapplied to MDEIT. The annular electrode mode makes the z direction current densityimage on the xy plane consistent with the corresponding conductivity image.Consequently, MDEIT can be simplified to current density imaging, shortening thedata measurement time and image reconstruction time.The paper introduces the current density based algorithm, the magnetic fluxdensity based algorithm and the differential imaging method. The inverse problem ofMDEIT is ill-posed. To solve the ill-posed inverse problem of MDEIT, a new imagereconstruction algorithm based on total variation (TV) regularization and a new hybrid reconstruction algorithm combined the TV regularization and L1normregularization for the sparse image are proposed. Compared with the L2normregularization, the TV regularization preserves local smoothness and piecewiseconstancy. The hybrid regularization simultaneously encourages properties of sparsityand smoothness in the reconstructed image, leading to the better location of the target.The image resolution and contrast in MDEIT image reconstruction are affectedby parameters such as the measurement arrangement. This paper firstly uses singularvalue analysis to obtain the superior measurement arrangement by comparingdifferent measurement configurations. Then the redundancy reduction is applied todiscard the measurement points having more correlated conductivity information toobtain the optimum measurement arrangement. The results indicate that properlyincreasing the number of current injections and the number of measurement circles,and locating preferentially the electrodes and detectors on the region nearest to theRoI produce more useful singular values and better reconstructed images.This paper designs a data acquisition system of MDEIT. The software isdeveloped based on LabVIEW, aiming for the automatic data collection and storageof many measurement points. Based on the data acquisition system, the experimentsare operated on the rod phantom, the saline-filled phantom and the ager phantom, withthe two types of electrode modes yielding transverse or annular injections. Thefeasibility of MDEIT is demonstrated with the reconstructed images from the physicalphantoms, advancing MDEIT towards a new medical imaging technique.