Carey M. Rappaport
Edwin Marengo, Jose Angel Martinez-Lorenzo
Date of Award
Doctor of Philosophy
Department or Academic Unit
College of Engineering. Department of Electrical and Computer Engineering.
Electrical engineering, Inverse problems, Inverse scattering, Optimization techniques
Tomography, Diagnostic imaging
Biomedical Engineering and Bioengineering
A novel method of optical diffraction tomography (ODT) to image weakly scattering, electrically large two dimensional (2D) objects using the far-zone scattered field data is presented. The proposed technique is based on the expansion of the target object function in terms of Fourier-Bessel basis functions, and an alternative approximation for the total electric field within the support of the investigated scatterer. Analytical (Mie) plane-wave scattering by a layered, circularly symmetric, lossy dielectric cylinder, and finite-difference time-domain simulations involving plane-wave scattering first by a more general, lossless dielectric phantom and then by embryo models involving various mitochondrial distributions are utilized to compare the performance of the proposed method with that of the standard ODT techniques which are based on the Born/Rytov approximations and Fourier diffraction theorem. Quantitative and qualitative superiority of the presented method is demonstrated. The proposed method can be used without being confined to far-zone observations with a proper (cylindrical-spherical) receiver configuration, and can be easily modified to handle multi-frequency data for a wideband reconstruction whenever applicable. The proposed 2D technique can be readily extended to more realistic three dimensional scenarios, and can be used in tomographic microscopy for various purposes, such as the cell-embryo health assessment for in vitro fertilization procedures.
Karbeyaz, Ersel, "Modal-based tomographic imaging from far-zone observations" (2009). Electrical Engineering Dissertations. Paper 13. http://hdl.handle.net/2047/d10019157
Click button above to open, or right-click to save.COinS