Clutter-based metrics for radar-based breast imaging: clutter region selection

Abstract

Results from first-in-human and later trials of microwave breast imaging continue to be reported with a variety of quantitative and qualitative findings in the literature. Clutter-based metrics are very widely reported, however, small differences in the implementations and definitions are common and restrict the ability to compare these values across findings. Using idealised images and experimental validation, two common versions of the signal-to-mean ratio (SMR) are compared and the maximum and typical differences in the final values examined. For the largest signal regions, difference of up to 10% could be expected, with the largest differences observed in homogeneous breast phantoms. Based on these findings, not including the signal region in the clutter region for SMR is recommended to match the only practical definition of the signal-to-clutter ratio (SCR).