{"id":3622,"date":"2023-12-20T01:54:11","date_gmt":"2023-12-20T06:54:11","guid":{"rendered":"https:\/\/my.vanderbilt.edu\/masi\/?p=3622"},"modified":"2024-01-23T16:38:44","modified_gmt":"2024-01-23T21:38:44","slug":"enhancing-hierarchical-transformers-for-whole-brain-segmentation-with-intracranial-measurements-integration","status":"publish","type":"post","link":"https:\/\/my.vanderbilt.edu\/masi\/2023\/12\/enhancing-hierarchical-transformers-for-whole-brain-segmentation-with-intracranial-measurements-integration\/","title":{"rendered":"Enhancing Hierarchical Transformers for Whole Brain Segmentation with Intracranial Measurements Integration"},"content":{"rendered":"<p>Xin Yu, Yucheng Tang, Qi Yang, Ho Hin Lee, Shunxing Bao, Yuankai Huo, and Bennett A. Landman. &#8220;Enhancing Hierarchical Transformers for Whole Brain Segmentation with Intracranial Measurements Integration.&#8221;\u00a0SPIE Medical Imaging 2024<\/p>\n<p><strong>Full text: NIHMSID<\/strong><\/p>\n<p><strong>Abstract:<\/strong><\/p>\n<p><i>Whole brain segmentation with magnetic resonance imaging (MRI) enables the non-invasive measurement of\u00a0<\/i><i>brain regions, including total intracranial volume (TICV) and posterior fossa volume (PFV). Enhancing the\u00a0<\/i><i>existing whole brain segmentation methodology to incorporate intracranial measurements offers a heightened\u00a0<\/i><i>level of comprehensiveness in the analysis of brain structures. Despite its potential, the task of generalizing deep\u00a0<\/i><i>learning techniques for intracranial measurements faces data availability constraints due to limited manually\u00a0<\/i><i>annotated atlases encompassing whole brain and TICV\/PFV labels. In this paper, we enhancing the hierarchical\u00a0<\/i><i>transformer UNesT for whole brain segmentation to achieve segmenting whole brain with 133 classes\u00a0<\/i><i>and TICV\/PFV simultaneously. To address the problem of data scarcity, the model is first pretrained on\u00a0<\/i><i>4859 T1-weighted (T1w) 3D volumes sourced from 8 different sites. These volumes are processed through a\u00a0<\/i><i>multi-atlas segmentation pipeline for label generation, while TICV\/PFV labels are unavailable. Subsequently,\u00a0<\/i><i>the model is finetuned with 45 T1w 3D volumes from Open Access Series Imaging Studies (OASIS) where\u00a0<\/i><i>both 133 whole brain classes and TICV\/PFV labels are available. We evaluate our method with Dice similarity\u00a0<\/i><i>coefficients(DSC). We show that our model is able to conduct precise TICV\/PFV estimation while maintaining\u00a0<\/i><i>the 132 brain regions performance at a comparable level. Code and trained model are available at:\u00a0<\/i><i>https:\/\/github.com\/MASILab\/UNesT\/wholebrainSeg<\/i><i>.<\/i><\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-medium wp-image-3624\" src=\"https:\/\/my.vanderbilt.edu\/masi\/wp-content\/uploads\/sites\/2304\n2661\/2023\/12\/vis_brain-1-300x237.png\" alt=\"vis_brain-1\" width=\"300\" height=\"237\" srcset=\"https:\/\/cdn.vanderbilt.edu\/t2-my\/my-prd\/wp-content\/uploads\/sites\/2304\/2023\/12\/vis_brain-1-300x237.png 300w, https:\/\/cdn.vanderbilt.edu\/t2-my\/my-prd\/wp-content\/uploads\/sites\/2304\/2023\/12\/vis_brain-1-768x607.png 768w, https:\/\/cdn.vanderbilt.edu\/t2-my\/my-prd\/wp-content\/uploads\/sites\/2304\/2023\/12\/vis_brain-1-650x514.png 650w, https:\/\/cdn.vanderbilt.edu\/t2-my\/my-prd\/wp-content\/uploads\/sites\/2304\/2023\/12\/vis_brain-1.png 1739w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Xin Yu, Yucheng Tang, Qi Yang, Ho Hin Lee, Shunxing Bao, Yuankai Huo, and Bennett A. Landman. &#8220;Enhancing Hierarchical Transformers for Whole Brain Segmentation with Intracranial Measurements Integration.&#8221;\u00a0SPIE Medical Imaging 2024 Full text: NIHMSID Abstract: Whole brain segmentation with magnetic resonance imaging (MRI) enables the non-invasive measurement of\u00a0brain regions, including total intracranial volume (TICV) and&#8230;<\/p>\n","protected":false},"author":8987,"featured_media":3624,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[138,130,3],"tags":[],"class_list":["post-3622","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-deep-brain-stimulation","category-deep-learning","category-image-segmentation"],"_links":{"self":[{"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/posts\/3622","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/users\/8987"}],"replies":[{"embeddable":true,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/comments?post=3622"}],"version-history":[{"count":3,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/posts\/3622\/revisions"}],"predecessor-version":[{"id":3627,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/posts\/3622\/revisions\/3627"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/media\/3624"}],"wp:attachment":[{"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/media?parent=3622"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/categories?post=3622"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/my.vanderbilt.edu\/masi\/wp-json\/wp\/v2\/tags?post=3622"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}