Real-time analysis of wide-field sub-diffuse images of normal and cancerous skin tissue using an artificial neural network

Simple item page
dc.contributor.advisorTunnell, James W.
dc.contributor.advisorMarkey, Mia Kathleen
dc.contributor.committeeMemberHollmig, Tyler
dc.contributor.committeeMemberCaramanis, Constantine
dc.contributor.committeeMemberPorter, Emily
dc.creatorStier, Andrew Clayton
dc.date.accessioned2021-07-27T00:25:37Z
dc.date.available2021-07-27T00:25:37Z
dc.date.created2021-05
dc.date.issued2021-05
dc.date.submittedMay 2021
dc.date.updated2021-07-27T00:25:38Z
dc.description.abstractSub-diffuse optical properties have the potential to serve as cancer biomarkers. Sub-diffuse spatial frequency domain imaging (sd-SFDI) can reveal wide-field heatmaps of these properties and has advantages over other imaging modalities. However, rendering optical property heatmaps from these images is currently too slow for this technology to be considered for real-time applications, such as image guided surgery. Moreover, the research on sub-diffuse optical properties and their utility for identifying cancerous tissue is limited. This dissertation demonstrates the rendering of sub-diffuse optical property heatmaps from experimental sd-SFDI images in real time and examines how these properties differ across normal and cancerous skin tissue subtypes. To achieve these results, a Monte Carlo model was developed with a novel phase function sampling method which enabled simulation of sd-SFDI spectra over the wide range of sub-diffuse optical properties found in biological tissue and calibration phantoms. Datasets simulated with this model were used to train an artificial neural network (ANN) machine learning model to predict sub-diffuse optical properties from sd-SFDI reflection spectra. The speed and accuracy of the ANN was tested on sd-SFDI images of tissue simulating phantoms. The ANN was then applied to sd-SFDI images of tissue samples taken from skin cancer patients to render optical property heatmaps of the samples. The sub-diffuse optical properties of basal cell carcinoma (BCC), adipose, and dermis regions within these samples, guided by co-registered and marked histology slides, were analyzed. The ANN accurately processed the sd-SFDI images and rendered optical property heatmaps from experimental data in real time. Quantitative differences were seen between the optical properties of the different tissue subtypes. These results bring the overall process of sd-SFDI a fundamental step closer to real-time speeds, advance the knowledge of sub-diffuse optical properties of biological tissue, and set a foundation for future real-time medical applications of sd-SFDI.
dc.description.departmentElectrical and Computer Engineering
dc.format.mimetypeapplication/pdf
dc.identifier.urihttps://hdl.handle.net/2152/86938
dc.identifier.urihttp://dx.doi.org/10.26153/tsw/13888
dc.language.isoen
dc.subjectLight scattering
dc.subjectOptical properties
dc.subjectTissue optics
dc.subjectMachine learning
dc.subjectArtificial neural network
dc.titleReal-time analysis of wide-field sub-diffuse images of normal and cancerous skin tissue using an artificial neural network
dc.typeThesis
dc.type.materialtext
thesis.degree.departmentElectrical and Computer Engineering
thesis.degree.disciplineElectrical and Computer Engineering
thesis.degree.grantorThe University of Texas at Austin
thesis.degree.levelDoctoral
thesis.degree.nameDoctor of Philosophy

Access full-text files

Original bundle

Now showing 1 - 1 of 1
Thumbnail Image
Name:
STIER-DISSERTATION-2021.pdf
Size:
5.71 MB
Format:
Adobe Portable Document Format
Download

License bundle

Now showing 1 - 2 of 2
No Thumbnail Available
Name:
PROQUEST_LICENSE.txt
Size:
4.45 KB
Format:
Plain Text
Description:
Download
No Thumbnail Available
Name:
LICENSE.txt
Size:
1.84 KB
Format:
Plain Text
Description:
Download

Collections

UT Electronic Theses and Dissertations