{"id":461,"date":"2025-12-27T12:11:52","date_gmt":"2025-12-27T15:11:52","guid":{"rendered":"https:\/\/wp.ufpel.edu.br\/gaia\/?p=461"},"modified":"2025-12-27T12:11:52","modified_gmt":"2025-12-27T15:11:52","slug":"ieramiars1","status":"publish","type":"post","link":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/","title":{"rendered":"Liver Tumor Segmentation in CT Scans Using Deep Learning: A U-Net Approach with Transfer Learning","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).<\/p>\n<p>Leia o artigo completo: <a href=\"https:\/\/doi.org\/10.5753\/eramiars.2025.16788\" target=\"_blank\">https:\/\/doi.org\/10.5753\/eramiars.2025.16788<\/a><\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":133,"featured_media":464,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"link","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[466],"tags":[469,472,474,473,468,467,471,470],"class_list":{"0":"post-461","1":"post","2":"type-post","3":"status-publish","4":"format-link","5":"has-post-thumbnail","6":"hentry","7":"category-ieramiars","8":"tag-deep-learning","9":"tag-inteligencia-artificial-em-radiologia","10":"tag-liver-tumor-segmentation","11":"tag-medical-image-segmentation","12":"tag-segmentacao-de-figado-em-tomografia-computadorizada","13":"tag-segmentacao-de-tumores-hepaticos","14":"tag-transfer-learning","15":"tag-u-net","16":"post_format-post-format-link","18":"post-with-thumbnail","19":"post-with-thumbnail-large"},"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.7 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Liver Tumor Segmentation in CT Scans Using Deep Learning<\/title>\n<meta name=\"description\" content=\"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\" \/>\n<meta property=\"og:locale\" content=\"pt_BR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Liver Tumor Segmentation in CT Scans Using Deep Learning\" \/>\n<meta property=\"og:description\" content=\"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).\" \/>\n<meta property=\"og:url\" content=\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\" \/>\n<meta property=\"og:site_name\" content=\"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial\" \/>\n<meta property=\"article:published_time\" content=\"2025-12-27T15:11:52+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/eramia-logo-horizontal-1.png\" \/>\n\t<meta property=\"og:image:width\" content=\"612\" \/>\n\t<meta property=\"og:image:height\" content=\"200\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"gaia\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:title\" content=\"Liver Tumor Segmentation in CT Scans Using Deep Learning\" \/>\n<meta name=\"twitter:description\" content=\"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/eramia-logo-horizontal-1.png?fit=612%2C200&ssl=1\" \/>\n<meta name=\"twitter:label1\" content=\"Escrito por\" \/>\n\t<meta name=\"twitter:data1\" content=\"gaia\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. tempo de leitura\" \/>\n\t<meta name=\"twitter:data2\" content=\"1 minuto\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"ScholarlyArticle\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\"},\"author\":{\"name\":\"gaia\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/700c500555f9133a6c44fd223e393507\"},\"headline\":\"Liver Tumor Segmentation in CT Scans Using Deep Learning: A U-Net Approach with Transfer Learning\",\"datePublished\":\"2025-12-27T15:11:52+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\"},\"wordCount\":133,\"publisher\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#organization\"},\"image\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1\",\"keywords\":[\"deep learning\",\"intelig\u00eancia artificial em radiologia\",\"liver tumor segmentation\",\"medical image segmentation\",\"segmenta\u00e7\u00e3o de f\u00edgado em tomografia computadorizada\",\"segmenta\u00e7\u00e3o de tumores hep\u00e1ticos\",\"transfer learning\",\"u-net\"],\"articleSection\":[\"I ERAMIA\u2011RS\"],\"inLanguage\":\"pt-BR\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\",\"url\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\",\"name\":\"Liver Tumor Segmentation in CT Scans Using Deep Learning\",\"isPartOf\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1\",\"datePublished\":\"2025-12-27T15:11:52+00:00\",\"description\":\"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).\",\"breadcrumb\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#breadcrumb\"},\"inLanguage\":\"pt-BR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-BR\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage\",\"url\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1\",\"contentUrl\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1\",\"width\":528,\"height\":517},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"In\u00edcio\",\"item\":\"https:\/\/wp.ufpel.edu.br\/gaia\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Liver Tumor Segmentation in CT Scans Using Deep Learning: A U-Net Approach with Transfer Learning\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#website\",\"url\":\"https:\/\/wp.ufpel.edu.br\/gaia\/\",\"name\":\"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial\",\"description\":\"Promovemos a investiga\u00e7\u00e3o cient\u00edfica te\u00f3rica e pr\u00e1tica sobre t\u00e9cnicas de IA e no emprego destas t\u00e9cnicas em tecnologias inovadoras.\",\"publisher\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/wp.ufpel.edu.br\/gaia\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"pt-BR\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#organization\",\"name\":\"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial\",\"url\":\"https:\/\/wp.ufpel.edu.br\/gaia\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-BR\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/11\/cropped-gaiaufpel_logo-1.jpeg?fit=200%2C148&ssl=1\",\"contentUrl\":\"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/11\/cropped-gaiaufpel_logo-1.jpeg?fit=200%2C148&ssl=1\",\"width\":200,\"height\":148,\"caption\":\"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial\"},\"image\":{\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.instagram.com\/gaia.ufpel\/\",\"https:\/\/www.linkedin.com\/company\/gaiaufpel\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/700c500555f9133a6c44fd223e393507\",\"name\":\"gaia\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-BR\",\"@id\":\"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/525fd84c409b2d96667f51e5da3c8858fdfa0eb1d38a6c8eb957f99fc519ed82?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/525fd84c409b2d96667f51e5da3c8858fdfa0eb1d38a6c8eb957f99fc519ed82?s=96&d=mm&r=g\",\"caption\":\"gaia\"},\"url\":\"https:\/\/wp.ufpel.edu.br\/gaia\/author\/gaia\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Liver Tumor Segmentation in CT Scans Using Deep Learning","description":"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/","og_locale":"pt_BR","og_type":"article","og_title":"Liver Tumor Segmentation in CT Scans Using Deep Learning","og_description":"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).","og_url":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/","og_site_name":"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial","article_published_time":"2025-12-27T15:11:52+00:00","og_image":[{"width":612,"height":200,"url":"https:\/\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/eramia-logo-horizontal-1.png","type":"image\/png"}],"author":"gaia","twitter_card":"summary_large_image","twitter_title":"Liver Tumor Segmentation in CT Scans Using Deep Learning","twitter_description":"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).","twitter_image":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/eramia-logo-horizontal-1.png?fit=612%2C200&ssl=1","twitter_misc":{"Escrito por":"gaia","Est. tempo de leitura":"1 minuto"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"ScholarlyArticle","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#article","isPartOf":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/"},"author":{"name":"gaia","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/700c500555f9133a6c44fd223e393507"},"headline":"Liver Tumor Segmentation in CT Scans Using Deep Learning: A U-Net Approach with Transfer Learning","datePublished":"2025-12-27T15:11:52+00:00","mainEntityOfPage":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/"},"wordCount":133,"publisher":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#organization"},"image":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1","keywords":["deep learning","intelig\u00eancia artificial em radiologia","liver tumor segmentation","medical image segmentation","segmenta\u00e7\u00e3o de f\u00edgado em tomografia computadorizada","segmenta\u00e7\u00e3o de tumores hep\u00e1ticos","transfer learning","u-net"],"articleSection":["I ERAMIA\u2011RS"],"inLanguage":"pt-BR"},{"@type":"WebPage","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/","url":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/","name":"Liver Tumor Segmentation in CT Scans Using Deep Learning","isPartOf":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#website"},"primaryImageOfPage":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage"},"image":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1","datePublished":"2025-12-27T15:11:52+00:00","description":"Este trabalho apresenta um pipeline para segmenta\u00e7\u00e3o 2D de f\u00edgado e tumores a partir de tomografia computadorizada (NIfTI). Utilizamos o dataset p\u00fablico Liver Tumor Segmentation com 131 volumes, do qual foram extra\u00eddas fatias 2D e convertidas em imagens JPEG (com compress\u00e3o m\u00ednima, sem perda percept\u00edvel) para treinamento da U-Net com backbone ResNet50 (transfer learning) implementando a biblioteca FastAI. O pr\u00e9-processamento inclui leitura dos volumes com NiBabel, aplica\u00e7\u00e3o de janelas DICOM espec\u00edficas para f\u00edgado, normaliza\u00e7\u00e3o e redimensionamento para 128\u00d7128. Como fun\u00e7\u00e3o de perda foi adotada CrossEntropy (multiclasse) e m\u00e9tricas customizadas de acur\u00e1cia sobre os pixels de interesse (foreground).","breadcrumb":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#breadcrumb"},"inLanguage":"pt-BR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/"]}]},{"@type":"ImageObject","inLanguage":"pt-BR","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#primaryimage","url":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1","contentUrl":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1","width":528,"height":517},{"@type":"BreadcrumbList","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/ieramiars1\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"In\u00edcio","item":"https:\/\/wp.ufpel.edu.br\/gaia\/"},{"@type":"ListItem","position":2,"name":"Liver Tumor Segmentation in CT Scans Using Deep Learning: A U-Net Approach with Transfer Learning"}]},{"@type":"WebSite","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#website","url":"https:\/\/wp.ufpel.edu.br\/gaia\/","name":"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial","description":"Promovemos a investiga\u00e7\u00e3o cient\u00edfica te\u00f3rica e pr\u00e1tica sobre t\u00e9cnicas de IA e no emprego destas t\u00e9cnicas em tecnologias inovadoras.","publisher":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/wp.ufpel.edu.br\/gaia\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"pt-BR"},{"@type":"Organization","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#organization","name":"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial","url":"https:\/\/wp.ufpel.edu.br\/gaia\/","logo":{"@type":"ImageObject","inLanguage":"pt-BR","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/logo\/image\/","url":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/11\/cropped-gaiaufpel_logo-1.jpeg?fit=200%2C148&ssl=1","contentUrl":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/11\/cropped-gaiaufpel_logo-1.jpeg?fit=200%2C148&ssl=1","width":200,"height":148,"caption":"Grupo de Aplica\u00e7\u00f5es em Intelig\u00eancia Artificial"},"image":{"@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.instagram.com\/gaia.ufpel\/","https:\/\/www.linkedin.com\/company\/gaiaufpel"]},{"@type":"Person","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/700c500555f9133a6c44fd223e393507","name":"gaia","image":{"@type":"ImageObject","inLanguage":"pt-BR","@id":"https:\/\/wp.ufpel.edu.br\/gaia\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/525fd84c409b2d96667f51e5da3c8858fdfa0eb1d38a6c8eb957f99fc519ed82?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/525fd84c409b2d96667f51e5da3c8858fdfa0eb1d38a6c8eb957f99fc519ed82?s=96&d=mm&r=g","caption":"gaia"},"url":"https:\/\/wp.ufpel.edu.br\/gaia\/author\/gaia\/"}]}},"jetpack_featured_media_url":"https:\/\/i0.wp.com\/wp.ufpel.edu.br\/gaia\/files\/2025\/12\/Liver-Tumor-Segmentation-in-CT-Scans-Using-Deep-Learning.png?fit=528%2C517&ssl=1","jetpack-related-posts":[{"id":195,"url":"https:\/\/wp.ufpel.edu.br\/gaia\/pequeno-tutorial-de-opencv-com-modulo-dnn-para-deep-learning\/","url_meta":{"origin":461,"position":0},"title":"Pequeno tutorial de OpenCV com m\u00f3dulo DNN para Deep Learning","author":"gaia","date":"15\/04\/2021","format":false,"excerpt":"https:\/\/learnopencv.com\/deep-learning-with-opencvs-dnn-module-a-definitive-guide\/?ck_subscriber_id=272173871#what-is-opencvv-dnn-module","rel":"","context":"Em &quot;Not\u00edcias&quot;","block_context":{"text":"Not\u00edcias","link":"https:\/\/wp.ufpel.edu.br\/gaia\/category\/noticias\/"},"img":{"alt_text":"","src":"","width":0,"height":0},"classes":[]},{"id":191,"url":"https:\/\/wp.ufpel.edu.br\/gaia\/o-curso-de-aprendizado-profundo-de-yann-lecun-esta-online-e-gratuito\/","url_meta":{"origin":461,"position":1},"title":"O curso de aprendizado profundo de Yann LeCun  est\u00e1 online e gratuito.","author":"gaia","date":"29\/03\/2021","format":false,"excerpt":"https:\/\/nyudatascience.medium.com\/yann-lecuns-deep-learning-course-at-cds-is-now-fully-online-accessible-to-all-787ddc8bf0af","rel":"","context":"Em &quot;Not\u00edcias&quot;","block_context":{"text":"Not\u00edcias","link":"https:\/\/wp.ufpel.edu.br\/gaia\/category\/noticias\/"},"img":{"alt_text":"","src":"","width":0,"height":0},"classes":[]}],"jetpack_sharing_enabled":true,"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/posts\/461","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/users\/133"}],"replies":[{"embeddable":true,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/comments?post=461"}],"version-history":[{"count":1,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/posts\/461\/revisions"}],"predecessor-version":[{"id":465,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/posts\/461\/revisions\/465"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/media\/464"}],"wp:attachment":[{"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/media?parent=461"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/categories?post=461"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/wp.ufpel.edu.br\/gaia\/wp-json\/wp\/v2\/tags?post=461"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}