Toward model-free and reference-free quantitative ultrasound

Nguyen, Trong Ngoc

Permalink

https://hdl.handle.net/2142/105027 Copy

Description

Title

Toward model-free and reference-free quantitative ultrasound

Author(s)

Nguyen, Trong Ngoc

Issue Date

2019-04-16

Director of Research (if dissertation) or Advisor (if thesis)

Oelze, Michael L.

Doctoral Committee Chair(s)

Oelze, Michael L.

Committee Member(s)

• Do, Minh N.
• Boppart, Stephen A.
• Liang, Zhi-Pei

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

Quantitative ultrasound, reference-free, fatty liver, in situ calibration

Abstract

Current spectral-based quantitative ultrasound (QUS) techniques rely on a reference or a calibration signal to obtain system-independent ultrasonic scattering parameters. Adoption of QUS clinically has been hindered by (1) the need to limit clinical ultrasonic scanner settings to account for a finite number of calibration signals or (2) the need to acquire calibration signals immediately before or after a setting is changed on an ultrasonic scanner, which can interrupt the workflow in the busy clinical environment. An additional factor that hinders the effectiveness of QUS clinically is the presence of layers over the tissue regions to be interrogated. Layer effects are not accounted for when using a reference phantom, reducing the technique's reliability. The dissertation presents experimental and computational approaches to improve the accuracy of QUS estimates in the presence of intervening layers using an in situ calibration target and the feasibility of reference-free quantitative ultrasound using a convolutional neural network. The sensitivity of CNN in a reference-free environment is assessed and experimentally validated. A model-free approach through the use of principal component analysis is proposed as a general method for tissue characterization using QUS and compared to a model-based approach. To address the effects of intervening layers on QUS accuracy and precision, an in situ calibration approach is proposed and experimentally verified to improve the QUS estimates.

Graduation Semester

2019-05

Type of Resource

text

Permalink

http://hdl.handle.net/2142/105027

Copyright and License Information

Copyright 2019 Trong Nguyen

Owning Collections