{"id":{"repo_id":"uoit","oai_identifier":"oai:ontariotechu.scholaris.ca:10155/2095"},"canonical_url":"https://search.dev.ndltd.org/etd/uoit/oai:ontariotechu.scholaris.ca:10155/2095","repository":{"repo_id":"uoit","name":"Ontario Institute of Technology","base_url":"https://ontariotechu.scholaris.ca/server/oai/request"},"display":{"title":"Uncertainty-aware fusion of foundation and task-specific models for cardiac MRI segmentation","abstract":"Vision foundation models, such as the Segment Anything Model (SAM), demonstrate strong zero-shot generalization but lack precision with anatomically challenging structures. In contrast, convolutional neural network (CNN)-based models achieve high accuracy on domain-specific data but struggle to generalize to unseen data. To address these complementary limitations, we propose an uncertainty-aware fusion framework that integrates the generalizability of foundation models with the anatomical precision of task-specific models for cardiac MRI segmentation. The approach combines Dempster-Shafer Theory (DST) with an entropy-guided fallback mechanism to perform voxel-wise fusion of calibrated probability maps. DST fusion is applied in regions of agreement, while high-conflict regions are handled by selecting predictions from the model with lower uncertainty. Experiments on in-domain and cross-domain datasets show consistent improvements, with larger gains under domain shift. To the best of our knowledge, this is the first voxel-wise uncertainty-based DST fusion of foundation and task-specific models for cardiac MRI segmentation.","abstract_html":"Vision foundation models, such as the Segment Anything Model (SAM), demonstrate strong zero-shot generalization but lack precision with anatomically challenging structures. In contrast, convolutional neural network (CNN)-based models achieve high accuracy on domain-specific data but struggle to generalize to unseen data. To address these complementary limitations, we propose an uncertainty-aware fusion framework that integrates the generalizability of foundation models with the anatomical precision of task-specific models for cardiac MRI segmentation. The approach combines Dempster-Shafer Theory (DST) with an entropy-guided fallback mechanism to perform voxel-wise fusion of calibrated probability maps. DST fusion is applied in regions of agreement, while high-conflict regions are handled by selecting predictions from the model with lower uncertainty. Experiments on in-domain and cross-domain datasets show consistent improvements, with larger gains under domain shift. To the best of our knowledge, this is the first voxel-wise uncertainty-based DST fusion of foundation and task-specific models for cardiac MRI segmentation.","abstract_has_math":false,"creators":["Rumman, Mosarrat"],"institution":"University of Ontario Institute of Technology","degree_name":"Master of Science (MSc)","degree_level":null,"degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":[],"advisors":["Ebrahimi, Mehran","Davoudi, Kourosh"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-04-01","date_published":"2026-04-01","updated_at":"2026-07-24T05:35:36Z","subjects":[],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10155/2095","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Ebrahimi, Mehran","Davoudi, Kourosh"]},{"key":"dc:creator","label":"Author","values":["Rumman, Mosarrat"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-04-28T19:45:17Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-04-01"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Master of Science (MSc)"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Ontario Institute of Technology"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10155/2095"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Vision foundation models, such as the Segment Anything Model (SAM), demonstrate strong zero-shot generalization but lack precision with anatomically challenging structures. In contrast, convolutional neural network (CNN)-based models achieve high accuracy on domain-specific data but struggle to generalize to unseen data. To address these complementary limitations, we propose an uncertainty-aware fusion framework that integrates the generalizability of foundation models with the anatomical precision of task-specific models for cardiac MRI segmentation. The approach combines Dempster-Shafer Theory (DST) with an entropy-guided fallback mechanism to perform voxel-wise fusion of calibrated probability maps. DST fusion is applied in regions of agreement, while high-conflict regions are handled by selecting predictions from the model with lower uncertainty. Experiments on in-domain and cross-domain datasets show consistent improvements, with larger gains under domain shift. To the best of our knowledge, this is the first voxel-wise uncertainty-based DST fusion of foundation and task-specific models for cardiac MRI segmentation."]},{"key":"dc:title","label":"Title","values":["Uncertainty-aware fusion of foundation and task-specific models for cardiac MRI segmentation"]}]}],"canonical_facts":{"dc:contributor.advisor":["Ebrahimi, Mehran","Davoudi, Kourosh"],"dc:creator":["Rumman, Mosarrat"],"dc:date.accessioned":["2026-04-28T19:45:17Z"],"dc:date.issued":["2026-04-01"],"dc:description.abstract":["Vision foundation models, such as the Segment Anything Model (SAM), demonstrate strong zero-shot generalization but lack precision with anatomically challenging structures. In contrast, convolutional neural network (CNN)-based models achieve high accuracy on domain-specific data but struggle to generalize to unseen data. To address these complementary limitations, we propose an uncertainty-aware fusion framework that integrates the generalizability of foundation models with the anatomical precision of task-specific models for cardiac MRI segmentation. The approach combines Dempster-Shafer Theory (DST) with an entropy-guided fallback mechanism to perform voxel-wise fusion of calibrated probability maps. DST fusion is applied in regions of agreement, while high-conflict regions are handled by selecting predictions from the model with lower uncertainty. Experiments on in-domain and cross-domain datasets show consistent improvements, with larger gains under domain shift. To the best of our knowledge, this is the first voxel-wise uncertainty-based DST fusion of foundation and task-specific models for cardiac MRI segmentation."],"dc:identifier.uri":["https://hdl.handle.net/10155/2095"],"dc:language.iso":["en"],"dc:title":["Uncertainty-aware fusion of foundation and task-specific models for cardiac MRI segmentation"],"dc:type":["Thesis"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_name":["Master of Science (MSc)"],"thesis:institution_name":["University of Ontario Institute of Technology"]},"updated_at":"2026-07-24T05:35:36Z"}