The image reconstruction is an ill-posed inverse problem of flnding such internal im- pedivity distribution that minimizes certain optimization criteria. The optimization necessitates algorithms that impose regularization and some prior information constraint. The regularization techniques vary in their complexity. This paper proposes new variants of the regularization tech- niques to be used for the acquirement of more accurate reconstruction results and the possibility of the applying difierential evolution algorithms in an optimization process. DOI: 10.2529/PIERS061006093457 The image reconstruction problem is a widely investigated problem with many applications in physical and biological sciences. The Electrical impedance tomography (EIT) can be used for reconstruction process. The theoretical background of EIT is given in (1). The currents are applied through the electrodes attached to the surface of the object and the resulting voltages are measured using the same or additional electrodes. Internal impedivity distribution is recalculated from the measured voltages and currents. It is well known that while the forward problem is well-posed, the inverse problem is highly ill-posed. Various numerical techniques with difierent advantages have been developed to solve this problem. The aim is to reconstruct, as accurately and fast as possible, the impedivity distribution in two or three dimensional models. Usually, a set of voltage measurements is acquired from the boundaries of the determined volume, whilst it is subjected to a sequence of low-frequency current patterns, which are preferred to direct current ones to avoid polarization efiects. Since the frequency of the injected current is su-ciently low, usually in the range of 10{100kHz, EIT can be treated as a quasi-static problem. So we only consider the conductivity for simplicity. The scalar potential U can be therefore introduced, and so the resulting fleld is conservative and the continuity equation for the volume current density can be expressed by the potential U div(ae gradU) = 0 (1) Equation (1) together with the modifled complete electrode model equations (2) are discretized by the flnite element method (FEM) in the usual way. Using FEM we calculate approximate values of electrode voltages for the approximate element conductivity vector ae (NE£1), NE is the number of flnite elements. Furthermore, we assume the constant approximation of a conductivity distribution ae on the flnite element region. The forward EIT calculation yields an estimation of the electric potential fleld in the interior of the volume under certain Neumann and Dirichlet boundary conditions. The FEM in two or three dimensions is exploited for the forward problem with current sources. 2. SOLUTION OF INVERSE PROBLEM Image reconstruction of EIT is an inverse problem, which is usually presented as minimizing the suitable objective function (ae) relative to ae: To minimize the objective function (ae) we can use a deterministic approach based on the Least Squares (LS) method. Due to the ill-posed nature of the problem, regularization has to be used. First the standard Tikhonov Regularization method (TRM) described in (3) was used to solve this inverse EIT problem