Publication: Microwave imaging systems using dasweighted phase coherent factor for breast cancer detection
Date
2023-04-01
Authors
Rasammal A/P Rasappan
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis describes an improved radar based imaging systemfor breast cancer
detection that employs 32 array p-slot ultra-wideband antennae. There are numerous
algorithms utilized in beamforming UltraWide-Band Imaging (UWBI) system and the
array processing algorithms is the one that focuses or steers the array in a particular
direction. Several popular algorithms utilized for beamforming UWBI systems
include DAS, DMAS, IDAS, CF-DAS, CR-DAS and RCB. Despite the latest
advancement in UWBI technology, a microwave imaging system is still subjected to
a limitation due to breast heterogeneity. UWBI method for breast cancer is based on
the significant contrast between the dielectric properties of the healthy and malignant
tissues, however recent studies concluded the contrast has dropped by 10% in
heterogeneity breast. In UWBI, limited resolution and cluttering and side lobes further
complicate the problem. The wavelength, λ for microwaves lies almost in the same
order with the length of the target of interest. Therefore, employing higher frequencies
to obtain better resolutions and improved imaging accuracy remains a challenge in
microwave imaging. The heavy cluttering due to heterogeneity of the breast and the
penetration depth further worsen the UWBI system. The side lobes and clutters affect
the quality of the images in terms of the resolution and sub-clutter visibility.
Therefore, this study attempts to address the problems by (i) to investigate an element
positions of a 3D antenna array to locate and differentiate tumours with higher
precision in heterogeneity breast, (ii) propose a side lobe suppression technique for the
conventional DAS beamformer in order to enhance the resolution and image quality and finally, (iii) validate and evaluate the performance of the proposed algorithm using
simulated and experimental data. The new reconstruction approach adds the Phase
Coherence Factor (PCF) into the traditional delay and sum (DAS) beamforming
algorithm, substantially eliminating side- and grating-lobe interference noise. Several
breast models made from chemical mixtures generated on the basis of realistic human
tissues are used to evaluate the system. Each model is housed in a hemispherical breast
radome made of polylactide and encircled by 32 p-slot antennae arranged in four
concentric layers. Two 16-channel multiplexers connect these antennae to an 8.5 GHz
vector network analyser, automatically switching different combinations of transmitter
and receiver pairs in a sequential way. With an average signal-to-clutter ratio of 7.0
dB and a full-width half-maximum of 2.3 mm, the system can accurately detect 5 mm
tumours in a complicated and homogeneously dense 3D breast model. As part of future
research, this study lays the path for a clinical trial involving human subjects.