Shear wave elastography is an ultrasound-based method that converts shear wave properties of the tissue of interest into tissue characteristics, in order to differentiate diseased tissue (stiffer) from healthy tissue. This relationship between shear wave propagation and material characteristics is known for large isotropic elastic media such as liver and breast tissue, but is difficult to theoretically derive for arterial or cardiac applications because of their complex geometry and material properties (bounded anisotropic hyper- and viscoelastic media). As plenty of shear wave simulations in arterial and/or cardiac phantoms are currently available at the bioMMeda-lab, the aim of this thesis is to develop a deep learning network trained on the available simulation data to predict material properties based on shear wave propagation patterns. Actual experimental data can then be used later on to validate the developed network.
This thesis aims at developing a deep learning network to estimate tissue material properties based on shear wave propagation patterns.