Advancing photoacoustic neuroimaging through deep learning
Kuo, Joseph
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/117691
Description
Title
Advancing photoacoustic neuroimaging through deep learning
Author(s)
Kuo, Joseph
Issue Date
2022-12-09
Director of Research (if dissertation) or Advisor (if thesis)
Anastasio, Mark A
Department of Study
Electrical & Computer Eng
Discipline
Electrical & Computer Engr
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
Photoacoustic
Deep Learning
Abstract
Photoacoustic computed tomography (PACT) is a promising brain imaging modality in which the optically induced initial pressure distribution is reconstructed from the measured ultrasonic wavefields. Unlike x-ray computed tomography, PACT exposes the patient to no ionizing radiation. Computationally efficient image reconstruction algorithms have been developed using a homogeneous acoustic medium. However, this assumption is unwarranted in brain imaging due to the elastic and acoustic heterogeneities of the skull. To compensate for these heterogeneities, wave equation-based reconstruction algorithms have been developed based on the elastic finite-difference time-domain method. These methods yield high-quality images if the elastic and acoustic properties of the skull are known exactly. However, model-based reconstruction algorithms are generally computationally burdensome, making them ill-suited for functional imaging. To address these issues, we propose a two-step 3D reconstruction algorithm. The first step uses a computationally efficient but approximate image reconstruction algorithm. In the second step, a high-quality image is obtained by removing aberrations from the previous step using a 3-D convolutional neural network. The proposed approach is validated on computed simulation studies and compared with traditional model-based approaches.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
