Estimation of Organ Boundaries in Electrical Impedance Tomography

In electrical impedance tomography (EIT) the impedance distribution is usually estimated in xed elements inside the object. The implicit assumption is most often that the impedance in each element is more or less independent of the other (neighboring) elements. In many cases, such as in the impedance imaging of the chest, this model might not be a feasible one. An example is the heart muscle and blood, whose impedances do not change but instead the associated volumes depend on the phase of the cardiac cycle. In this paper we propose a method in which the internal organ boundaries are estimated instead of the impedances in some element. The method is based on expressing the boundaries as truncated radial Fourier series and the estimation of the Fourier coeecients with the aid of the associated linearized mapping and the nite element method. The feasibility of the method is shown by tank measurements.