Research On Magnetic Resonance Electrical Impedance Tomography

With the development of medical imaging technology,magnetic resonance electrical impedance tomography(MREIT)has become a hot research field in biomedical engineering.MREIT is a new generation of impedance imaging technology combined with electrical impedance tomography(EIT)technology and magnetic resonance current density imaging(MRCDI)technology.According to the principle that the conductivity of biological tissues changes greatly before the lesions of the generator,MREIT technology can detect the lesions in time during the incubation period or early stage of the disease,which greatly advances the diagnosis time,improves the success rate of treatment,the research on this technology has broad prospects and important significance.The MREIT electrical impedance imaging problem can be summarized as the following two parts: forward problem and inverse problem.The current common practice for forward problem solving is through finite element analysis.In the solution of the inverse problem,in addition to some traditional algorithms,in recent years,researchers have gradually applied some intelligent algorithms to image reconstruction,in which differential evolution(DE)algorithm is used as a robust and simple global optimization evolutionary algorithm has shown good performance in MREIT,which has attracted the attention of many researchers.However,due to its own principles,the algorithm also has some inherent problems in dealing with MREIT problems.Based on the DE-based MREIT image reconstruction algorithm,this paper innovates and improves the shortcomings of the algorithm.The main research contents are as follows:1?In the DE-based MREIT image reconstruction algorithm,the MREIT forward problem is solved throughout the entire process of the algorithm,while the forward problem solving is computationally intensive and time consuming,which leads to the lower overall efficiency of the algorithm.In order to solve this problem,the paper proposes an MREIT algorithm based on RBF neural network and DE idea.The RBF neural network is used to construct the nonlinear function relationship between resistivity and objective function to replace the traditional call forward problem for individual evaluation,which reduces the pair.The number of calls to the forward problem greatly improves the efficiency of the algorithm under the premise of error.Simulation experiments verify the effectiveness of the proposed algorithm.2?The DE-based MREIT image reconstruction algorithm needs to evaluate the objective function for all population individuals every time,and there are problems such as low search efficiency and population aggregation phenomenon.By introducing the simplex search method to the modified DE algorithm,the paper ensures the diversity of the DE algorithm and accelerates the search speed and search accuracy of the algorithm.The simulation results show that the improved algorithm can obtain a clear impedance reconstruction image.Compared with the basic DE algorithm,the improved image reconstruc tion error and convergence speed are significantly improved,which effectively improves the overall performance of the algorithm.