Optimisation of Artefact Removal Algorithm for Microwave Imaging of the Axillary Region Using Experimental Prototype Signals

Abstract

Microwave Imaging (MWI) has the potential to aid breast cancer staging through the detection of Axillary Lymph Nodes (ALNs). This type of system can present some challenges, mainly due to the irregular axillary surface. The optimisation of the artefact removal algorithm to successfully remove the surface reflections is of great importance. In this paper, we propose using Singular Value Decomposition (SVD) as an artefact removal algorithm and study the effect of choosing different subsets of antenna positions for artefact removal on imaging results using experimental signals. We show that different subsets of antenna positions affect the results and in some cases prevent the targets detection. Our analysis allowed us to find an optimal combination of parameters which results in Signal-to-Clutter Ratio higher than 2.77 dB and Location Error lower than 14.9 mm for three different experimental tests. These results are relevant for the development of dedicated algorithms for ALN-MWI application.

Evaluating the Performance of Algorithms in Axillary Microwave Imaging towards Improved Breast Cancer Staging

“…Moreover, considering the complexity of the axillary region shape, two main novelties were employed in this study: we considered a higher frequency band with a central frequency of 7.5 GHz when compared to a central frequency of 4 GHz in [ 16 , 18 , 44 ]; and we evaluated a planar antenna positioning spread over a distance of 80 mm when compared to a cylindrical antenna positioning [ 16 , 18 ]. A higher frequency band allows for reduced antenna dimensions and an increase in the density of antenna positions spread over the axillary region.…”