摘要：AbstractElectrical Impedance Tomography (EIT) is a radiation-free imaging technique. The reconstructed EIT image represents the conductivity distribution changes of the human body. EIT images are reconstructed based on the Finite Element Method (FEM). Currently, the 3D EIT reconstruction is desired. However, because of the huge amount of parameters associated with the FEM model, the 3D EIT reconstructions are neither stable nor computationally efficient. In this paper, we present a classification-reconstruction method for EIT imaging. In the classification step, based on a fast initial linear reconstruction, the FEM elements are grouped into several clusters. The reconstructions are based on these clusters by forcing the conductivity distribution the same within each cluster. The proposed framework is efficient in computation and potentially fit for 3D EIT reconstruction.
关键词：KeywordsElectrical Impedance TomographyFinite Element MethodConditional Random Fields
