{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/121317"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/121317","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Atlas driven low-rank reconstruction of dynamic speech magnetic resonance imaging","abstract":"Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2025-08-01","abstract_html":"Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2025-08-01","abstract_has_math":false,"creators":["Jin, Riwei"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Bioengineering","degree_department":null,"school":null,"contributors":["Sutton, Brad","Liang, Zhi-Pei","Shosted, Ryan","Lam, Fan"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2023,"date_issued":"2023-08","date_published":"2023-08","updated_at":"2026-07-22T22:24:57Z","subjects":["Dynamic Mri","Speech Imaging","Low-rank Reconstruction","Atlasing"],"languages":["en","eng"],"rights":["Copyright 2023 Riwei Jin"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2142/121317","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Sutton, Brad","Liang, Zhi-Pei","Shosted, Ryan","Lam, Fan"]},{"key":"dc:creator","label":"Author","values":["Jin, Riwei"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2023-08","2023-07-05"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Bioengineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Dynamic Mri","Speech Imaging","Low-rank Reconstruction","Atlasing"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en","eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2023 Riwei Jin"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://hdl.handle.net/2142/121317"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2025-08-01","The student, Riwei Jin, accepted the attached license on 2023-06-23 at 16:08.","The student, Riwei Jin, submitted this Dissertation for approval on 2023-06-23 at 16:18.","This Dissertation was approved for publication on 2023-07-05 at 16:36.","DSpace SAF Submission Ingestion Package generated from Vireo submission #19456 on 2023-12-04 at 17:18:05","Dynamic magnetic resonance imaging (MRI) offers great potential for speech-related studies and clinical applications, as it enables the simultaneous investigation of velopharyngeal motion and physiological properties in real time. However, the practical applications of dynamic speech MRI are often constrained by various technical tradeoffs, such as imaging speed, spatial coverage, spatial resolution, acquisition time, and clinical interpretation. One significant limitation is that long acquisitions are not acceptable for children under ten years old, which necessitates a reduction in acquisition time while maintaining high spatiotemporal resolution and full vocal tract coverage. Furthermore, there is a requirement for a clinical tool to analyze motion variations between individual subjects and the average population, specifically for studying the velopharyngeal dysfunction (VPD) of cleft palate patients. To address these challenges, this dissertation proposes an atlas-driven residual reconstruction approach with a further optimized dynamic speech imaging methodology which achieves a spatial resolution of 2 mm isotropic, a temporal resolution of 40 frames per second (FPS), a full-vocal-tract coverage with 32 slices, and a short acquisition time of less than 10 minutes. For the 10-slice children's protocol, the acquisition time can be further reduced to 2.5 minutes. This dissertation introduces a Partial Separability (PS) model-based imaging protocol that incorporates several optimizations to improve the efficiency and quality of the imaging process. These optimizations include: 1. Introducing an interlaced sampling pattern for imaging and navigator datasets, which allows for separate handling of the spatial and temporal subspaces. 2. Implementing a sparse temporal sampling navigator technique that reduces the overall acquisition time by 40%. 3. Utilizing a Poisson-disc random under-sampling approach in both the phase and slice directions, effectively reducing sampling coherence. 4. Employing a locally higher-rank operator that specifically targets the oropharyngeal region of interest. 5. Developing a regional-optimized temporal basis that, when combined with the locally higher-rank operator, enhances the temporal signal-to-noise ratio (SNR) and the reconstruction quality. 6. Implementing an optimization algorithm based on Alternating Direction Method of Multipliers (ADMM) for one-norm regularization, which aids in the reconstruction process. We used the time strip plots, subtraction analysis, quiver plots of deformation fields from resting to sample states to show the capability that the method can achieve. The atlas-driven reconstruction approach was further introduced by leveraging a spatiotemporal atlas from specific speech samples from 20 subjects. To implement this method within individual subjects, a temporal and spatial alignment process was necessary, aligning the atlas space to the subject space. By utilizing this approach, the dissertation showcased the potential and capability of enhancing reconstruction quality and visualizing the variations between individual subjects and the average population. This was demonstrated through the presentation of timeframe images and time strip plots, effectively demonstrating the improvements made. However, it is important to note that there are still areas that warrant further investigation and improvement in future research to achieve high quality spatiotemporal images of speech in even shorter acquisition times."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Atlas driven low-rank reconstruction of dynamic speech magnetic resonance imaging"]}]}],"canonical_facts":{"dc:contributor":["Sutton, Brad","Liang, Zhi-Pei","Shosted, Ryan","Lam, Fan"],"dc:creator":["Jin, Riwei"],"dc:date":["2023-08","2023-07-05"],"dc:description":["Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2025-08-01","The student, Riwei Jin, accepted the attached license on 2023-06-23 at 16:08.","The student, Riwei Jin, submitted this Dissertation for approval on 2023-06-23 at 16:18.","This Dissertation was approved for publication on 2023-07-05 at 16:36.","DSpace SAF Submission Ingestion Package generated from Vireo submission #19456 on 2023-12-04 at 17:18:05","Dynamic magnetic resonance imaging (MRI) offers great potential for speech-related studies and clinical applications, as it enables the simultaneous investigation of velopharyngeal motion and physiological properties in real time. However, the practical applications of dynamic speech MRI are often constrained by various technical tradeoffs, such as imaging speed, spatial coverage, spatial resolution, acquisition time, and clinical interpretation. One significant limitation is that long acquisitions are not acceptable for children under ten years old, which necessitates a reduction in acquisition time while maintaining high spatiotemporal resolution and full vocal tract coverage. Furthermore, there is a requirement for a clinical tool to analyze motion variations between individual subjects and the average population, specifically for studying the velopharyngeal dysfunction (VPD) of cleft palate patients. To address these challenges, this dissertation proposes an atlas-driven residual reconstruction approach with a further optimized dynamic speech imaging methodology which achieves a spatial resolution of 2 mm isotropic, a temporal resolution of 40 frames per second (FPS), a full-vocal-tract coverage with 32 slices, and a short acquisition time of less than 10 minutes. For the 10-slice children's protocol, the acquisition time can be further reduced to 2.5 minutes. This dissertation introduces a Partial Separability (PS) model-based imaging protocol that incorporates several optimizations to improve the efficiency and quality of the imaging process. These optimizations include: 1. Introducing an interlaced sampling pattern for imaging and navigator datasets, which allows for separate handling of the spatial and temporal subspaces. 2. Implementing a sparse temporal sampling navigator technique that reduces the overall acquisition time by 40%. 3. Utilizing a Poisson-disc random under-sampling approach in both the phase and slice directions, effectively reducing sampling coherence. 4. Employing a locally higher-rank operator that specifically targets the oropharyngeal region of interest. 5. Developing a regional-optimized temporal basis that, when combined with the locally higher-rank operator, enhances the temporal signal-to-noise ratio (SNR) and the reconstruction quality. 6. Implementing an optimization algorithm based on Alternating Direction Method of Multipliers (ADMM) for one-norm regularization, which aids in the reconstruction process. We used the time strip plots, subtraction analysis, quiver plots of deformation fields from resting to sample states to show the capability that the method can achieve. The atlas-driven reconstruction approach was further introduced by leveraging a spatiotemporal atlas from specific speech samples from 20 subjects. To implement this method within individual subjects, a temporal and spatial alignment process was necessary, aligning the atlas space to the subject space. By utilizing this approach, the dissertation showcased the potential and capability of enhancing reconstruction quality and visualizing the variations between individual subjects and the average population. This was demonstrated through the presentation of timeframe images and time strip plots, effectively demonstrating the improvements made. However, it is important to note that there are still areas that warrant further investigation and improvement in future research to achieve high quality spatiotemporal images of speech in even shorter acquisition times."],"dc:format":["application/pdf"],"dc:identifier":["https://hdl.handle.net/2142/121317"],"dc:language":["en","eng"],"dc:rights":["Copyright 2023 Riwei Jin"],"dc:subject":["Dynamic Mri","Speech Imaging","Low-rank Reconstruction","Atlasing"],"dc:title":["Atlas driven low-rank reconstruction of dynamic speech magnetic resonance imaging"],"dc:type":["text"],"thesis:degree_discipline":["Bioengineering"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:57Z"}