Segmentacija krvnih žila u slikama očne pozadine

undergraduate thesis

• Title (english): Segmentation of Blood Vessels in the Images of the Ocular Background
• Author: Jakov Kruljac
• Mentor: Marko Subašić (mentor)
• Committee member: Marko Subašić (predsjednik povjerenstva)
• Committee member: Ivana Matovinović (član povjerenstva)
• Committee member: Denis Milošević (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom radu obrađena je tema segmentiranja krvnih žila iz očne pozadine. Objašnjene su tri vrste dubokih neuronskih mreža koje su potrebe za rješavanje problema: potpuno povezana neuronska mreža, konvolucijska neuronska mreža i generativna kontradiktorna mreža. Posebno je objašnjena vrsta mreže koja se koristi kao generator u generativnoj kontradiktornoj neuronskoj mreži, U-net. Za diskriminator se koristi se obična neuronska mreža koja podatke uspoređuje na temelju cijele slike. Model generativne neuronske mreže koji je napravljen u sklopu ovog rada poprilično je uspješan u segmentaciji krvnih žila, a rezultati se mogu vidjeti u radu. Za model koristio se programski jezik Python 3.6 te dodatne biblioteke namijenjene za duboko učenje.

Abstract (english)

This paper deals with the topic of segmentation of blood vessels from the fundus. Three types of deep neural networks that are needed to solve the problem are explained: a fully connected neuronal network, a convolutional neuron network, and a generative adversarial network. Particularly explained is the type of network used as a generator in the generative adversarial neural network U-net. A discriminator uses a regular neural network that compares data across the entire image. The model of generative adversarial neural network made in this work is quite successful in the segmentation of blood vessels, and the results can be seen in the work. The model used the Python 3.6 programming language and additional deep-learning libraries.

• Keywords: očna pozadina; segmentacija krvnih žila; duboke neuronske mreže; konvolucijske neuronske mreže; generativne kontradiktorne neuronske mreže; U-net; TensorFlow
• Keywords (english): fundus; blood vessel segmentation; deep neural network; convolution neural network; generative adversarial neural network; U-net; TensorFlow
• Language: croatian
• URN:NBN: urn:nbn:hr:168:164266
• 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: 2022-12-28 09:11:24