A clustering based dual model framework for EIT imaging: first experimental results

Abstract

Abstract Electrical Impedance tomography (EIT) imaging suffers greatly from the illposedness of the corresponding inverse problem. This is mainly caused by the high degree of freedom and the relatively large noise. One attempt to circumvent these difficulties is to use dual models. This article introduces a clustering based non-uniform dual model construction. With this framework, finite elements are grouped to reduce the complexity in inverse computations. The simulation and experiment results indicated that the k-means clustering method did not only preserve the sharp variations over conductivity mediums but also greatly filtered out artefacts found in the standard approach.