UWB Microwave Breast Cancer Detection: Generalized Models and Performance Prediction

Abstract

This paper presents a generic framework for the modeling of ultra-wideband (UWB) signal propagation in human breast, which facilitates system-level simulations and provides performance prediction. The clutter associated with the breast tissue heterogeneity is modeled through several key parameters depending on the tissue compositions. Subsequently, important channel properties such as the backscatter energy and the probability density function of time-of-arrival are derived. The modified Hermite polynomials, which fit well into the real pulse shapes, are then used to model the UWB signals. Armed with the channel/signal model preliminaries, three metrics are proposed, namely, the mean clutter response, the clean tumor response, and the worst-case clutter response. The generalized model provides a parsimonious way to study the effects of tissue structures, pulse templates, and array setup on the performance of a specified UWB imaging system. Numerical examples are used to demonstrate the usefulness of the proposed approach.