A Breast Cancer Detection System Using Metasurfaces With a Convolution Neural Network: A Feasibility Study

This work presents a breast cancer detection modality using a metasurface as the imaging medium and a microwave source. The fundamental concept is to record the electromagnetic energy, which, when transmitted through the breast, scatters differently according to the variation in the electrical properties of the breast tissue. The energy captured by the metasurface forms an impression image of the breast under test. This impression is used as a detection tool to determine the presence of a tumor inside the breast. A convolution neural network (CNN) is applied to enhance detection accuracy and provide quantitative data about tumors when present. The CNN results show the possibility of finding a tumor with a 10 mm diameter and distinguish between healthy and unhealthy breasts with 93% accuracy while determining the size and the size with location with an accuracy of 91% and 74%, respectively.