Segmentacija srca i krvnih žila iz medicinskih slika korištenjem duboke neuronske mreže

undergraduate thesis

• Title (english): Heart and Blood Vessel Segmentation From Medical Images Using Deep Neural Networks
• Author: Zrinka Pećanić
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (predsjednik povjerenstva)
• Committee member: Duje Medak (član povjerenstva)
• Committee member: Luka Posilović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

S bolestima krvožilnog sustava kao vodećim uzrokom smrti u svijetu logično je zapitati se može li tehnologija ubrzati/olakšati dijagnostiku i/ili liječenje. Cilj rada jest segmentirati sinkrotronske slike srca i krvnih žila kako bi se omogućilo brže prepoznavanje struktura (srčano tkivo, krvne žile) na istima. U radu je objašnjen korišteni model mreže U-Net – arhitektura i implementacija, hiperparametri te načini evaluacije i optimizacije modela. Predložene i implementirane su dvije vrste učenja: nadzirano – korištenjem samo označenih podataka u treniranju i polunadzirano – korištenjem označenih i neoznačenih podataka u treniranju.

Abstract (english)

With heart and blood vessel diseases being the leading cause of death globally, it is only logical to wonder how can technology help improve/ease their diagnostics and/or treatment. The goal was to create a model that will aid in recognizing structures (such as heart muscle tissue and blood vessels) using images obtained from the imaging machine synchrotron. In this paper we explain the U-Net model we used – its architecture and implementation, hyperparameters, as well as methods of model quality evaluation and optimization. Two network learning methods are proposed: supervised learning – exclusively using labeled data to train the model and semi-supervised learning – using both labeled and unlabeled data to train the model.

• Keywords: segmentacija; u-net; srce; žile; model; nadzirano; polunadzirano; neuron; učenje
• Keywords (english): segmentation; u-net; heart; vessels; model; supervised; semi-supervised; neuron; learning
• Language: croatian
• URN:NBN: urn:nbn:hr:168:053836
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 21:24:12