Generating a Library of Head Phantoms for Microwave Imaging using Spherical Harmonic Approximation

A novel method for generating a library of numerical head phantoms is presented. The parent model is created by parametrizing head tissue boundaries defined in the NEVA woman's model using spherical harmonic approximation. The parent model was then modified by stochastically altering the expansion coefficients. Electromagnetic simulations of the parent and modified model showed a notable difference in the scattering parameters. A large dataset essential for implementing microwave imaging algorithms based on artificial intelligence can be created using the proposed technique.