Joint Inversion of Electrical Impedance, Microwave and Ultrasonic Data With Structural Feature Fusion for Human Thorax Imaging

In this study, we propose a novel joint inversion method for human thorax imaging that combines electrical impedance tomography (EIT), microwave tomography (MT), and ultrasound tomography (UT) through structural feature fusion. The structural and value features of the three modality images are decoupled and encoded into structure code and value code by a disentangled variational autoencoder (DVAE). The inversion is then performed at the feature level, with the structure code and value code treated as unknown parameters to be inverted under a deterministic framework by the Gauss-Newton method. To ensure the consistency of the structures in the three modalities, the structure codes of the three modalities are averaged every few iterations during the inversion process. Numerical results demonstrate an improved performance of our joint inversion approach over traditional separate inversion methods in human thorax imaging.