Date of Award

5-2022

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mathematical Sciences

First Advisor

Vladislav Bukshtynov

Second Advisor

Christopher Bashur

Third Advisor

Ryan T. White

Fourth Advisor

Gnana Bhaskar Tenali

Abstract

A fully developed computational framework for the optimal reconstruction of binary-type images suitable for various models seen in biological and medical applications is developed and validated. This framework enables solutions to the inverse electrical impedance tomography (EIT) problems of cancer detection at different levels of complexity with multiple cancer-affected regions of different sizes based on available measurements usually affected by noise. A new spatial partitioning methodology and efficient scheme for switching between fine and coarse scales are developed to allow higher variations in the geometry of reconstructed binary images with superior performance confirmed computationally on various models. A nominal number of input parameters makes the approach simple for practical implementation in diverse settings. An easy-to-follow design of the entire framework allows extending its functionality to new models in many other applications and further enhancing its performance. The complete computational framework is tested in applications to 2D inverse EIT problems and demonstrates its high potential for improving the overall quality of EITbased procedures.

Included in

Mathematics Commons

Share

COinS