Spatial Prior for Quantitative Breast Cancer Microwave Imaging: a Comparison Between Non-Iterative Eigenfunction-Based Inversion and Sampling Methods

Breast cancer imaging represents a relevant and promising application area of microwave imaging technology. As breast tissues exhibit different electromagnetic properties at microwave frequencies depending on their typology and physio-pathological status, retrieving the complex-valued permittivity allows differentiating normal and diseased tissues. Recently, efforts have been made to obtain high-quality prior information in order to improve both the accuracy and resolution of microwave imaging reconstructions. In this work, the use of prior information based on a combination of linear/orthogonal sampling method results is proposed in order to generate spatial priors to be used within the Contrast Source Inversion scheme. Results are compared to contrast source inversion images generated using prior information obtain from a non-iterative eigenfunction-based inversion method.