BREAST CANCER DETECTION USING A HYBRID FINITE DIFFERENCE FREQUENCY DOMAIN AND PARTICLE SWARM OPTIMIZATION TECHNIQUES

Abstract
A hybrid technique based on Finite-difference frequency domain and particle swarm optimization techniques is proposed to reconstruct the breast cancer cell dimension and determines its position. Finite-difference frequency domain is formulated to calculate the scattered field after illuminating the breast by a microwave transmitter. Two-dimensional and three-dimensional models for the breast are used. The models include randomly distributed fatty breast tissue, glandular tissue, 2-mm thick skin, as well as chest wall tissue. The models are characterized by the dielectric properties of the normal breast tissue and malignant tissue at 800 MHz. Computer simulations have been performed by means of a numerical program; results show the capabilities of the proposed approach.