Files in this item

FilesDescriptionFormat

application/pdf

application/pdfNAGABHUSHANA-THESIS-2021.pdf (14MB)Restricted Access
(no description provided)PDF

Description

Title:Simulating acoustic wave reflection, transmission, and scattering in k-Wave for quantitative ultrasound applications
Author(s):Nagabhushana, Karthik Jedikere
Advisor(s):Han, Aiguo
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):acoustic simulation
acoustic modeling
k-Wave
quantitative ultrasound
modeling acoustic reflection and transmission
modeling acoustic scattering
modeling acoustic 3D impedance maps
Abstract:K-Wave is a versatile and powerful acoustic simulator that has gained popularity in the Medical Ultrasound research community. In this work, k-Wave was used to model two quantitative ultrasound (QUS) applications. In Part 1, detailed in Chapter 3, acoustic planar reflection and transmission were simulated to validate a new phantom attenuation measurement method. Attenuation coefficient (AC) measurements of reference phantoms, a key step in quantitative ultrasound, are complicated by transmission loss of laminate membranes. Conventional through-transmission methods overcome this issue by characterizing separately membrane and phantom material specimens. A simpler alternative that uses a single phantom to simultaneously measure the membrane transmission loss and phantom material AC is proposed. The proposed method was validated in simulation using the k-Wave toolbox. The acquired AC, between 0.5-1 dB/cm-MHz, had a maximum error of 0.06 dB/cm-MHz. The method was also experimentally validated wherein AC measurements were performed by two operators on five distinct phantoms, across five transducers, using the conventional method and the proposed method. The acquired AC, between 0.28-1.48 dB/cm-MHz, had a maximum error of 0.045 dB/cm-MHz across all phantoms. In Part 2, detailed in Chapters 4-6, acoustic scattering was simulated in k-Wave to extract backscattering parameters. This exercise was intended to be a precursor to a simulation of tissue histology-based computational phantoms known as 3D impedance maps (3DZMs) in k-Wave. In the past, 3DZM simulations have either used the spatial FFT approach or semi-analytical tools such as Field-II. However, both methods assume plane wave incidence, weak scattering, and the absence of multiple scattering in their analysis which is a limitation. A simulation in k-Wave transcends these assumptions and hence fulfills the full potential of a 3DZM. First, three schemes–focused transducer, near-field, and far-field were implemented in k-Wave to extract the differential backscatter cross-section of a 1-mm single fluid spherical scatterer. Planar reference method [1] with a single element spherically focused transducer was implemented in the focused transducer scheme. In the near-field scheme, the single scatterer was placed in the near-field of a wide planar transmitter and a point receiver was placed in the far-field. In the far-field scheme, the single scatterer was placed in the far-field of a piston transmitter and a point receiver was placed in the far-field. For a spatial resolution of 25 µm and within a frequency bandwidth of 1 to 3.5 MHz, the extracted differential backscatter cross-section’s root mean square (RMS) error was within 12.2%, 3.2%, and 3.4% of the peak value for the focused transducer, the near-field, and the far-field schemes respectively. Next, the backscatter coefficient (BSC) of a sparsely distributed collection of identical 20 µm fluid spherical scatterers was extracted in a radially averaged manner from a 192 µm wide spherical scattering volume. The extracted BSC was found to be a product of the single scatterer BSC and the number of scatterers in the same volume, thus matching the incoherent scattering theory. Finally, a staircase-free representation of the non-rectangular sources from the literature was successfully extended to heterogeneous simulation media, both for the single scatterer and the scatterer collection. For a spatial resolution of 100 µm and within a frequency bandwidth of 1 to 3.5 MHz, the extracted differential backscatter cross-section of a single 1 mm fluid spherical scatterer had an RMS error of 90.1% of the peak value. In contrast, the staircase-free version had an RMS error of 10.8% of the peak value. Overall, this work successfully set up a simulation flow to model backscattering in k-Wave, which can be extended to 3DZMs.
Issue Date:2021-12-10
Type:Thesis
URI:http://hdl.handle.net/2142/114114
Rights Information:Copyright 2021 Karthik Nagabhushana
Date Available in IDEALS:2022-04-29
Date Deposited:2021-12


This item appears in the following Collection(s)

Item Statistics