%A HUANG Ningning (黄宁宁), MA Yixin (马艺馨), ZHANG Mingzhu (张明珠), GE Hao (葛浩), WU Huawei (吴华伟) %T Finite Element Modeling of Human Thorax Based on MRI Images for EIT Image Reconstruction %0 Journal Article %D 2021 %J J Shanghai Jiaotong Univ Sci %R 10.1007/s12204-020-2232-x %P 33-39 %V 26 %N 1 %U {https://xuebao.sjtu.edu.cn/sjtu_en/CN/abstract/article_43115.shtml} %8 2021-02-28 %X Electrical impedance tomography (EIT) image reconstruction is a non-linear problem. In general, finite element model is the critical basis of EIT image reconstruction. A 3D human thorax modeling method for EIT image reconstruction is proposed herein to improve the accuracy and reduce the complexity of existing finite element modeling methods. The contours of human thorax and lungs are extracted from the layers of magnetic resonance imaging (MRI) images by an optimized Otsu’s method for the construction of the 3D human thorax model including the lung models. Furthermore, the GMSH tool is used for finite element subdivision to generate the 3D finite element model of human thorax. The proposed modeling method is fast and accurate, and it is universal for different types of MRI images. The effectiveness of the proposed method is validated by extensive numerical simulation in MATLAB. The results show that the individually oriented 3D finite element model can improve the reconstruction quality of the EIT images more effectively than the cylindrical model, the 2.5D model and other human chest models.