<?xml version="1.0" encoding="UTF-8"?><?xml-stylesheet type="text/xsl" href="static/style.xsl"?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-11-21T04:41:13Z</responseDate><request verb="GetRecord" identifier="oai:macsphere.mcmaster.ca:11375/32532" metadataPrefix="dim">https://prod-ms-be.lib.mcmaster.ca/server/oai/request</request><GetRecord><record><header><identifier>oai:macsphere.mcmaster.ca:11375/32532</identifier><datestamp>2026-09-24T06:20:11Z</datestamp><setSpec>com_11375_271</setSpec><setSpec>col_11375_272</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:doc="http://www.lyncode.com/xoai" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
   <dim:field mdschema="dc" element="contributor" qualifier="advisor">Noseworthy, Michael</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="advisor">Doyle, Thomas</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="author">Zhu, Calvin</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="department" lang="en_US">Biomedical Engineering</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="accessioned">2025-10-16T18:30:44Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="available">2025-10-16T18:30:44Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="issued">2025</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">http://hdl.handle.net/11375/32532</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">https://doi.org/10.71548/8823</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="abstract" lang="en_US">Significant progress in deep learning has positioned highly capable, fully automated&#xd;
medical imaging tools on the near horizon. However, translation from research into&#xd;
clinical practice is slower than the rapid technological advancements. There is an alternative&#xd;
that is more feasible and leverages strengths from both a human operator and&#xd;
machine learning based tools where a decision support system is in charge of augmenting&#xd;
human ability. Termed human autonomy teaming (HAT), the goal is to have teams of&#xd;
human operators with autonomous machines work together on required tasks. The adoption&#xd;
of HATs has unfortunately also been slow. Interestingly, transformer-based deep&#xd;
learning architectures bring unique properties, the attention mechanism and generative&#xd;
pre-training, that may have applications across different medical imaging related tasks&#xd;
and further investigation may provide insights for better adoption of HATs.&#xd;
Conceptually, the attention mechanism informs the model to pay attention to certain&#xd;
portions of the input. Using this as a means to examine model decisions, the attention&#xd;
heads of a Vision Transformer and the important pixels for a ResNet50 model were compared&#xd;
to radiologist annotations to determine appropriateness of each model’s attention&#xd;
mechanisms in classifying two datasets. Both the ResNet50 model and the ViT model&#xd;
had high levels of agreement with the radiologist annotations at 88.07% and 94.85%&#xd;
and 94.72% and 96.96% respectively; however, the vision transformers had better performance&#xd;
in both test accuracy and agreement.&#xd;
Generative pre-training is a method for training model weights without extra labelling.&#xd;
During training, samples have portions masked before being given to an autoencoder&#xd;
which tries to recover the original input. This training method has applications in&#xd;
areas of medical imaging such as reconstruction and classification on imbalanced data.&#xd;
Adapting generative pre-training methods for reconstructing undersampled k-space data&#xd;
with a 25% masking produced the most faithful reconstructions with an average MSE&#xd;
of 6.763 E-4 and an SSIM of 0.917. Interestingly, our results are more aligned with the&#xd;
optimal masking ratio of language models which have an ideal masking ratio between&#xd;
15-25% rather than the 75% masking ratio demonstrated in an imaging application. This&#xd;
reinforces the idea that further study is required to properly implement broader deep&#xd;
learning advances from computer vision in medical imaging. On the other hand, generative&#xd;
pre-training through masked autoencoders (MAEs) was investigated as a method of&#xd;
training on imbalanced data, achieving high single neuroimaging modality classification&#xd;
performance of 95.24%.&#xd;
Just as convolutional neural networks (CNNs) revolutionized the discipline, transformers&#xd;
have likewise established themselves as a defining breakthrough. Bringing many&#xd;
interesting avenues for exploration with them, transformers have pushed the state-of-theart&#xd;
and deep learning models now see use all around the world. Beyond the performance&#xd;
increases that transformers brought about, there is another lesson to be learned. Whatever&#xd;
the next ground-breaking architecture may be, researchers should consider delving into all aspects and applications rather than blindly hoping for performance improvements&#xd;
and fully automated tools.</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="degreetype" lang="en_US">Thesis</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="degree" lang="en_US">Doctor of Philosophy (PhD)</dim:field>
   <dim:field mdschema="dc" element="language" qualifier="iso" lang="en_US">en</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Deep Learning</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Human Autonomy Teaming</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Transformers</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Convolutional Neural Networks</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Machine Learning</dim:field>
   <dim:field mdschema="dc" element="title" lang="en_US">Enabling Human–Autonomy Teaming in Medical Imaging Through Transformers</dim:field>
   <dim:field mdschema="dc" element="type" lang="en_US">Thesis</dim:field>
   <dim:field mdschema="others" element="access-status">embargo</dim:field>
</dim:dim></metadata></record></GetRecord></OAI-PMH>