Novel 2D intelligent-decision genetic-algorithm based microwave imaging algorithm for breast cancer detection

The problem of non-invasive reconstruction of the internal structure of a dielectric cylinder had a great attention because of its usefulness in many applications such as medical computed tomography, underground targets imaging, microwave remote sensing, etc. Genetic algorithm (GA) based methodology is a candidate approach to solve the inverse portion of this microwave imaging problem. GA-based methodology is powerful and robust because of its independence of the initial guess and its few constrains on the solution domain, however, its main drawback is its overall run time. In this paper we focus on the breast cancer detection application where a very efficient GA-based microwave imaging algorithm is proposed and shown to be very accurate and fast. The proposed methodology is based on a novel two-steps GA-approach, where malignant and normal tissues are classified first in an approximate way, then their properties are refined in the second step. The performance of this proposed algorithm is shown to be around 50 times better than the regular GA-based imaging algorithm.