Breast Phantom Imaging Using Iteratively Corrected Coherence Factor Delay and Sum

Abstract

In this paper, a system for microwave breast tumor detection is presented using iteratively corrected coherence factor delay and sum (CF-DAS) algorithm. CF-DAS is data independent, which makes it stable in a noisy environment. However, data adaptive techniques have made significant progress by enhancing the image quality in microwave tomography. Thus, a novel data adaptive iterative variant of CF-DAS is proposed in this paper to produce stable and accurate images. The microwave imaging (MI) system contains a rotatable array of nine modified antipodal Vivaldi antennas in a circular arrangement, an array-mounting stand based on the stepper motor, the flexible phantom mounting podium, a control system for RF switching of the transceivers, and signal processing unit based on personal computer involved in the reconstruction of the image. The impedance bandwidth of the modified antenna is recorded from 2.5 to 11 GHz with stable directional radiation pattern. For performing the transmission and reception of the microwave signals, an SP8T nine port RF switch is used ranging from 2.5 to 8.0 GHz, and the switching is controlled by MATLAB software. Several low-cost lab-based homogenous and heterogeneous phantoms containing the dielectric property of human breast and tumor tissue are prepared to test the system efficiency. Since typical data independent radar-based techniques are ill-equipped for multiple reflection scenarios, an iteratively corrected variant of CF-DAS algorithm is used for processing the recorded backscattered signals to reconstruct the image of the breast phantom and to identify the existence and locate the area of the multiple breast tumors. The proposed method achieves more than 10-dB improvement over conventional CF-DAS in terms of signal to mean ratio for four different phantoms measured in this study.