Segmentation of fetal cerebral MRI using deep neural networks
Visualitza/Obre
Estadístiques de LA Referencia / Recolecta
Inclou dades d'ús des de 2022
Cita com:
hdl:2117/340254
Realitzat a/ambIMT Atlantique
Tipus de documentProjecte Final de Màster Oficial
Data2020-10
Condicions d'accésAccés obert
Tots els drets reservats. Aquesta obra està protegida pels drets de propietat intel·lectual i
industrial corresponents. Sense perjudici de les exempcions legals existents, queda prohibida la seva
reproducció, distribució, comunicació pública o transformació sense l'autorització del titular dels drets
Abstract
Fetal magnetic resonance imaging (MRI) is used for monitoring and characterizing fetal brain development from the 18th gestational week to term since MRI, with its superior soft tissue contrast resolution, provides much richer details compared with ultrasound images and any other imaging technique. The segmentation of the fetal brain MRIs is essential for their study. On one hand, segmentation of the fetal brain into different tissue classes is the key point of volumetric and morphological analysis in fetal MRI. On the other hand, localizing the fetal brain and obtaining a segmented mask to exclude the surrounding tissues is crucial to achieve accurate motion correction. The need for brain masks has motivated a series of studies into fetal brain extraction into fetal brain MRI. Cortical plate (CP) segmentation on fetal MRI is particularly challenging as the fetal CP is a very thin ribbon with a thickness that is comparable to the best achievable resolution on fetal MRI scans. Another factor that makes automatic segmentation of the fetal CP challenging is the substantial variations in fetal brain morphology due to the rapid development of the brain throughout gestation. Deep learning methods have often outperformed traditional machine learning and model-based methods in medical image analysis. One of the reasons for their successful implementations is their ability to extract the features relevant for the tasks directly from the data. The networks learn themselves to extract and interpret features relevant to the segmentation task. The Convolutional Neural Networks (CNN), a class of deep neural networks, is mostly applied to analyzing visual imagery since they are space invariant neural networks, which makes them ideal to work with images.
MatèriesImaging systems in medicine, Neural networks (Computer science), Imatgeria mèdica, Xarxes neuronals (Informàtica)
TitulacióMÀSTER UNIVERSITARI EN ENGINYERIA DE TELECOMUNICACIÓ (Pla 2013)
Col·leccions
Fitxers | Descripció | Mida | Format | Visualitza |
---|---|---|---|---|
TFM_Report_MRieraMarin.pdf | 3,044Mb | Visualitza/Obre |