Duboko učenje za segmentaciju krvnih žila srca

master's thesis

• Title (english): Deep Learning for Cardiac Blood Vessel Segmentation
• Author: Lucija Šošić
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (član povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Metode za segmentaciju medicinskih slika igraju kritičnu ulogu pri kvalitetnoj primjeni terapije, ali se pri tome moraju nositi s različitim ograničenjima koje ona povlači. U okviru diplomskog rada prikazan je pristup segmentaciji krvnih žila srca uporabom modela dubokog učenja. Detaljno je objašnjen cjevovod model gdje Mask R-CNN model vrši detekciju krvnih žila na slici, a potom UNet model daje konačnu semantičku masku. Rad je takod̄er posvećen opisu tehnika obrade slika kao odgovor na veliku složenost problema segmentacije medicinskih slika. Uspored̄eni su rezultati dobiveni uporabom različitih funkcija gubitka.

Abstract (english)

Methods for medical image segmentation play a critical role when it comes to the quality application of therapy, but they must deal with the various limitations that medical images entail. The thesis presents an approach to the segmentation of cardiac blood vessels using deep learning models. The pipeline model consists of the Mask R-CNN model that performs the detection of blood vessels in the image, and the UNet model that provides the final semantic mask. The paper is also devoted to the description of image processing techniques in response to the great complexity of medical image segmentation. The results obtained using different loss functions were compared.

• Keywords: računalni vid; segmentacija medicinskih slika; konvolucijske neuronske mreže; Mask R-CNN; UNet; cjevovod arhitektura
• Keywords (english): Computer Vision; Medical Image Segmentation; Convolutional Neural Networks; Mask R-CNN; UNet; Pipeline Architecture
• Language: croatian
• URN:NBN: urn:nbn:hr:168:657141
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 13:58:14