AI-Enhanced SDR Microwave Breast Cancer Screening Transceiver: A Comparative Study

Microwave transceivers play a crucial role in microwave breast screening systems, significantly influencing the final product performance, cost, and form factor. This article presents a portable and economically viable yet high-performing microwave transceiver designed based on the software-defined radio (SDR) architecture. This transceiver is intended to complement a switching matrix and antenna array, forming a novel microwave breast cancer screening prototype. This new system aims to replace two older transceivers, which are included in this study for comparative analysis. The comparison is conducted on a purposely designed test bench against a commercially available vector network analyzer (VNA) used as the benchmark. In addition, this VNA serves as a reference for training a denoising autoencoder neural network, aimed at enhancing the signal integrity of the new prototype. Comparative evaluations are performed across the time domain, frequency domain, and delay-multiply-and-sum (DMAS) image domain. The findings reveal that the SDR-based prototype achieves satisfactory and comparable imaging results, all while costing a fraction of the price of alternative systems.