Blending unsupervised and supervised mimic learning for discovering interpretable phenotypes in 3D imaging volumes

Simple item page
dc.contributor.authorNorouzi, Mahboobeh
dc.contributor.examiningcommitteeKordi, Behzad (Electrical and Computer Engineering)
dc.contributor.examiningcommitteeDavie, James (Biochemistry and Medical Genetics)
dc.contributor.supervisorAshraf, Ahmed
dc.contributor.supervisorKhan, Shehroz
dc.date.accessioned2023-08-22T17:12:04Z
dc.date.available2023-08-22T17:12:04Z
dc.date.issued2023-08-14
dc.date.submitted2023-08-14T21:49:15Zen_US
dc.degree.disciplineElectrical and Computer Engineeringen_US
dc.degree.levelMaster of Science (M.Sc.)
dc.description.abstractThe ability to differentiate between different subtypes of diseases is essential for informed clinical decision-making. However, in cases where only partial knowledge of a disease exists and its subtypes are unknown, traditional supervised classification approaches are not applicable. Instead, unsupervised methods are necessary for subtype discovery. Previous research in imaging-based subtype discovery has been limited, often lacking interpretability in the discovered phenotypes. This thesis proposes a novel data-driven method for discovering interpretable imaging phenotypes in 3D image volumes. While previous research has been limited, especially for 3D imaging modalities like MRI and CT, our approach focuses on CT scan images as a demonstration of 3D imaging. Our method combines unsupervised learning and supervised mimic learning for phenotype discovery and interpretability, respectively. We utilize unsupervised 3D autoencoders to discover phenotypes and employ supervised mimic learning to interpret our unsupervised pipeline. Notably, this is the first application of supervised mimic learning to understand an unsupervised model. Additionally, we introduce and formulate VolPAM, a technique enabling 3D interpretation of the discovered phenotypes. Our method is applicable to various medical imaging modalities and holds the potential to advance our understanding of respective diseases and conditions.
dc.description.noteOctober 2023
dc.description.sponsorshipUniversity of Manitoba Startup Fund
dc.identifier.urihttp://hdl.handle.net/1993/37479
dc.language.isoeng
dc.rightsopen accessen_US
dc.subjectUnsupervised Phenotype Discovery
dc.subjectInterpretable Deep Learning
dc.subjectBiomedical Imaging
dc.subjectArtificial Intelligence
dc.titleBlending unsupervised and supervised mimic learning for discovering interpretable phenotypes in 3D imaging volumes
dc.typemaster thesisen_US
local.subject.manitobano
oaire.awardNumberRGPIN/6457-2020
oaire.awardTitleNSERC Discovery Grant
project.funder.identifierhttp://dx.doi.org/10.13039/501100000038
project.funder.nameNatural Sciences and Engineering Research Council of Canada (NSERC)
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Thesis
Size:
7.35 MB
Format:
Adobe Portable Document Format
Description:
A 24 month embargo has been applied to meet the publisher conditions of deposit.
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
770 B
Format:
Item-specific license agreed to upon submission
Description:
Download
Collections
FGS - Electronic Theses and Practica