Array Sensitivity for Model-Based Microwave Breast Imaging

The success of microwave imaging via model-based inverse scattering is dependent on the quality of the model of the environment surrounding the imaging domain. Uncertainties in the properties of the environment may be significant enough to create a level of model-based noise that competes with the signals that carry information about the object's features of interest. In this paper, we investigate the impact of environmental material properties uncertainty on the problem of microwave breast imaging. Specifically, we analyze sensitivity to errors in the modeled environment, relative to sensitivity to breast tissue features of interest for a fixed enclosed array design. We also evaluate the impact of changing the array configuration (decreasing antenna-to-breast spacing and increasing breast compression) on these sensitivities. The results illustrate several regimes in which the model noise associated with uncertainty in the material properties in the array environment is comparable to the signals associated with the features of interest in the breast. These results highlight the importance of considering the effects of material properties uncertainty when evaluating reconstruction algorithms and array designs for microwave inverse scattering.