Microwave Radar-Based Breast Cancer Detection: Imaging in Inhomogeneous Breast Phantoms

Abstract

This letter presents, for the first time, experimental work on microwave breast cancer imaging using inhomogeneous breast phantoms. A recently designed 31-antenna array is used in imaging experiments. The imaging system operates in the full ultrawideband frequency range, between 3 and 10 GHz. To verify imaging performance of our system, new breast phantoms with inhomogeneous interior were developed. For three different breast phantoms presented in this work, the contrast between spherical phantom tumors and surrounding materials ranges from 5:1 to 1.6:1. Our results show that the biggest challenge in radar microwave imaging is the inhomogeneity of the volume being sensed, and not the contrast itself. In addition to experimental results, we also present the new image formation algorithm, which is a modified version of the delay-and-sum (DAS) algorithm. The new algorithm makes use of a new weighting factor, the coherence factor. The new algorithm is effective in reducing clutter, providing better images. For the most demanding imaging example presented herein, the new algorithm improves the peak clutter-to-target energy ratio by 3.1 dB.