Date of Graduation
Doctor of Philosophy in Engineering (PhD)
Second Committee Member
Third Committee Member
Breast Cancer, Image Segmentation, Machine Learning, Statistical Inference, Terahertz Imaging
Breast conserving surgery (BCS) is a common breast cancer treatment option, in which the cancerous tissue is excised while leaving most of the healthy breast tissue intact. The lack of in-situ margin evaluation unfortunately results in a re-excision rate of 20-30% for this type of procedure. This study aims to design statistical and machine learning segmentation algorithms for the detection of breast cancer in BCS by using terahertz (THz) imaging. Given the material characterization properties of the non-ionizing radiation in the THz range, we intend to employ the responses from the THz system to identify healthy and cancerous breast tissue in BCS samples. In particular, this dissertation covers the description of four segmentation algorithms for the detection of breast cancer in THz imaging. We first explore the performance of one-dimensional (1D) Gaussian mixture and t-mixture models with Markov chain Monte Carlo (MCMC). Second, we propose a novel low-dimension ordered orthogonal projection (LOOP) algorithm for the dimension reduction of the THz information through a modified Gram-Schmidt process. Once the key features within the THz waveform have been detected by LOOP, the segmentation algorithm employs a multivariate Gaussian mixture model with MCMC and expectation maximization (EM). Third, we explore the spatial information of each pixel within the THz image through a Markov random field (MRF) approach. Finally, we introduce a supervised multinomial probit regression algorithm with polynomial and kernel data representations. For evaluation purposes, this study makes use of fresh and formalin-fixed paraffin-embedded (FFPE) heterogeneous human and mice tissue models for the quantitative assessment of the segmentation performance in terms of receiver operating characteristics (ROC) curves. Overall, the experimental results demonstrate that the proposed approaches represent a promising technique for tissue segmentation within THz images of freshly excised breast cancer samples.
Chavez Esparza, T. (2021). Statistical Machine Learning for Breast Cancer Detection with Terahertz Imaging. Graduate Theses and Dissertations Retrieved from https://scholarworks.uark.edu/etd/3994