Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation
Uložené v:
| Názov: | Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation |
|---|---|
| Autori: | Kushibar, Kaisar, Salem, Mostafa, Rovira Cañellas, Alex, Salvi, Joaquim, Oliver, Arnau, Valverde, Sergi, Llado, Xavier |
| Prispievatelia: | Institut Català de la Salut, Kushibar K, Valverde S, Salvi J, Oliver A, Lladó X Institute of Computer Vision and Robotics, University of Girona, Girona, Spain. Salem M Institute of Computer Vision and Robotics, University of Girona, Girona, Spain. Computer Science Department, Faculty of Computers and Information, Assiut University, Asyut, Egypt. Rovira À Unitat de Ressonància Magnètica, Servei de Radiologia, Vall d'Hebron Hospital Universitari, Barcelona, Spain, Vall d'Hebron Barcelona Hospital Campus |
| Zdroj: | Scientia |
| Informácie o vydavateľovi: | Frontiers Media |
| Rok vydania: | 2021 |
| Predmety: | Cervell - Imatgeria per ressonància magnètica, Imatgeria (Tècnica), ANATOMY::Nervous System::Central Nervous System::Brain, Other subheadings::Other subheadings::Other subheadings::/diagnostic imaging, ANALYTICAL, DIAGNOSTIC AND THERAPEUTIC TECHNIQUES, AND EQUIPMENT::Diagnosis::Diagnostic Techniques and Procedures::Diagnostic Imaging::Tomography::Magnetic Resonance Imaging, ANATOMÍA::sistema nervioso::sistema nervioso central::encéfalo, Otros calificadores::Otros calificadores::Otros calificadores::/diagnóstico por imagen, TÉCNICAS Y EQUIPOS ANALÍTICOS, DIAGNÓSTICOS Y TERAPÉUTICOS::diagnóstico::técnicas y procedimientos diagnósticos::diagnóstico por imagen::tomografía::imagen por resonancia magnética |
| Popis: | Cervell; Imatge per ressonància magnètica; Aprenentatge transductiu ; Cerebro; Imagen de resonancia magnética; Aprendizaje transductivo ; Brain; Magnetic resonance imaging; Transductive learning ; Segmentation of brain images from Magnetic Resonance Images (MRI) is an indispensable step in clinical practice. Morphological changes of sub-cortical brain structures and quantification of brain lesions are considered biomarkers of neurological and neurodegenerative disorders and used for diagnosis, treatment planning, and monitoring disease progression. In recent years, deep learning methods showed an outstanding performance in medical image segmentation. However, these methods suffer from generalisability problem due to inter-centre and inter-scanner variabilities of the MRI images. The main objective of the study is to develop an automated deep learning segmentation approach that is accurate and robust to the variabilities in scanner and acquisition protocols. In this paper, we propose a transductive transfer learning approach for domain adaptation to reduce the domain-shift effect in brain MRI segmentation. The transductive scenario assumes that there are sets of images from two different domains: (1) source—images with manually annotated labels; and (2) target—images without expert annotations. Then, the network is jointly optimised integrating both source and target images into the transductive training process to segment the regions of interest and to minimise the domain-shift effect. We proposed to use a histogram loss in the feature level to carry out the latter optimisation problem. In order to demonstrate the benefit of the proposed approach, the method has been tested in two different brain MRI image segmentation problems using multi-centre and multi-scanner databases for: (1) sub-cortical brain structure segmentation; and (2) white matter hyperintensities segmentation. The experiments showed that the segmentation performance of a pre-trained model could be significantly improved by up to 10%. For the ... |
| Druh dokumentu: | article in journal/newspaper |
| Popis súboru: | application/pdf |
| Jazyk: | English |
| Relation: | Frontiers in Neuroscience;15; https://doi.org/10.3389/fnins.2021.608808; info:eu-repo/grantAgreement/ES/PE2013-2016/DPI2017-86696-R; Kushibar K, Salem M, Valverde S, Rovira À, Salvi J, Oliver A, et al. Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation. Front Neurosci. 2021 Apr 29;15:608808.; https://hdl.handle.net/11351/6704; 000649783300001 |
| DOI: | 10.3389/fnins.2021.608808 |
| Dostupnosť: | https://hdl.handle.net/11351/6704 https://doi.org/10.3389/fnins.2021.608808 |
| Rights: | Attribution 4.0 International ; http://creativecommons.org/licenses/by/4.0/ ; info:eu-repo/semantics/openAccess |
| Prístupové číslo: | edsbas.4845362C |
| Databáza: | BASE |
Nájsť tento článok vo Web of Science | Abstrakt: | Cervell; Imatge per ressonància magnètica; Aprenentatge transductiu ; Cerebro; Imagen de resonancia magnética; Aprendizaje transductivo ; Brain; Magnetic resonance imaging; Transductive learning ; Segmentation of brain images from Magnetic Resonance Images (MRI) is an indispensable step in clinical practice. Morphological changes of sub-cortical brain structures and quantification of brain lesions are considered biomarkers of neurological and neurodegenerative disorders and used for diagnosis, treatment planning, and monitoring disease progression. In recent years, deep learning methods showed an outstanding performance in medical image segmentation. However, these methods suffer from generalisability problem due to inter-centre and inter-scanner variabilities of the MRI images. The main objective of the study is to develop an automated deep learning segmentation approach that is accurate and robust to the variabilities in scanner and acquisition protocols. In this paper, we propose a transductive transfer learning approach for domain adaptation to reduce the domain-shift effect in brain MRI segmentation. The transductive scenario assumes that there are sets of images from two different domains: (1) source—images with manually annotated labels; and (2) target—images without expert annotations. Then, the network is jointly optimised integrating both source and target images into the transductive training process to segment the regions of interest and to minimise the domain-shift effect. We proposed to use a histogram loss in the feature level to carry out the latter optimisation problem. In order to demonstrate the benefit of the proposed approach, the method has been tested in two different brain MRI image segmentation problems using multi-centre and multi-scanner databases for: (1) sub-cortical brain structure segmentation; and (2) white matter hyperintensities segmentation. The experiments showed that the segmentation performance of a pre-trained model could be significantly improved by up to 10%. For the ... |
|---|---|
| DOI: | 10.3389/fnins.2021.608808 |