dc.contributor.advisor | Engan, Kjersti | |
dc.contributor.advisor | Kurz, Kathinka Dæhli | |
dc.contributor.author | Tomasetti, Luca | |
dc.date.accessioned | 2019-10-07T07:44:25Z | |
dc.date.available | 2019-10-07T07:44:25Z | |
dc.date.issued | 2019-06 | |
dc.identifier.uri | http://hdl.handle.net/11250/2620505 | |
dc.description | Master's thesis in Computer Science | nb_NO |
dc.description.abstract | This thesis explores different Convolutional Neural Network (CNN) approaches to classify and segment infarcted regions from images taken through a Computed Tomography Perfusion (CTP) from patients of the Stavanger’s hospital (SUS) affected by an ischemic stroke. Also, it evaluates the accuracy and the loss functions of the images analyzed through CNN. Furthermore, a segmentation approach, based on a U-Net model, is tested to create, from scratch, a unique image containing a summary of the section of the brain investigated with the different infarcted regions prediction. The purpose of this thesis work is to find a fast and effective method to help doctors in their decisions during these delicate and problematic situations. | nb_NO |
dc.language.iso | eng | nb_NO |
dc.publisher | University of Stavanger, Norway | nb_NO |
dc.relation.ispartofseries | Masteroppgave/UIS-TN-IDE/2019; | |
dc.subject | informasjonsteknologi | nb_NO |
dc.subject | datateknikk | nb_NO |
dc.subject | ischemic stroke | nb_NO |
dc.subject | image segmentation | nb_NO |
dc.subject | maskinlæring | nb_NO |
dc.subject | machine learning | nb_NO |
dc.subject | deep neural network | nb_NO |
dc.subject | convolutional neural network | nb_NO |
dc.subject | perfusion CT | nb_NO |
dc.title | Segmentation of infarcted regions in Perfusion CT images by 3D deep learning | nb_NO |
dc.type | Master thesis | nb_NO |
dc.description.version | updatedVersion | nb_NO |
dc.subject.nsi | VDP::Mathematics and natural science: 400::Information and communication science: 420 | nb_NO |