prikaz prve stranice dokumenta Prijenos obilježja slike primjenom konvolucijske neuronske mreže
  Download
PDF 4.83 MB
master's thesis
Prijenos obilježja slike primjenom konvolucijske neuronske mreže
2022. urn:nbn:hr:200:030103
Šarčević, Matej
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Software Engineering
Chair of Visual Computing

Cite this document

Šarčević, M. (2022). Prijenos obilježja slike primjenom konvolucijske neuronske mreže (Master's thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:030103

Šarčević, Matej. "Prijenos obilježja slike primjenom konvolucijske neuronske mreže." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2022. https://urn.nsk.hr/urn:nbn:hr:200:030103

Šarčević, Matej. "Prijenos obilježja slike primjenom konvolucijske neuronske mreže." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2022. https://urn.nsk.hr/urn:nbn:hr:200:030103

Šarčević, M. (2022). 'Prijenos obilježja slike primjenom konvolucijske neuronske mreže', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 21 March 2023, https://urn.nsk.hr/urn:nbn:hr:200:030103

Šarčević M. Prijenos obilježja slike primjenom konvolucijske neuronske mreže [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2022 [cited 2023 March 21] Available at: https://urn.nsk.hr/urn:nbn:hr:200:030103

M. Šarčević, "Prijenos obilježja slike primjenom konvolucijske neuronske mreže", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2022. Available at: https://urn.nsk.hr/urn:nbn:hr:200:030103

Please login to the repository to save this object to your list.
Metadata
TitlePrijenos obilježja slike primjenom konvolucijske neuronske mreže
Title (english)Transmission of image features using a convolutional neural network
AuthorMatej Šarčević
MentorČaslav Livada (mentor)
Committee memberČaslav Livada (predsjednik povjerenstva)
Committee memberAlfonzo Baumgartner (član povjerenstva)
Committee memberTomislav Galba (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Software Engineering)
(Chair of Visual Computing)
Osijek
Defense date and country2022-09-15, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
U ovom diplomskom radu opisane su osnove konvolucijskih neuronskih mreža. Kroz rad su opisane osnove i način rada neuronskih mreža, objašnjeni su sastavni dijelovi, odnosno slojevi mreže, te su predstavljeni primjeri najčešće korištenih konvolucijskih neuronskih mreža. Cilj praktičnog dijela rada je primjena konvolucijske neuronske mreže u postupku neuronskog prijenosa stila. To je postupak kojim se obilježja jedne slike prenose na drugu sliku. Opisana je teorijska strana i formule algoritma koje se koriste u algoritmu za neuronski prijenos stila. Programsko rješenje za navedeni postupak napisano je u programskom jeziku Python, točnije PyTorch biblioteci koja se koristi za razvoj i treniranje neuronskih mreža baziranih na dubokom učenju modela. Arhitektura konvolucijske neuronske mreže koja je korištena u rješenju je VGG19. Rezultati programskog rješenja prezentirani su u zadnjem dijelu rada gdje se može vidjeti kako izgleda generirana slika nakon prijenosa obilježja slike s jedne na drugu i kako kompleksnost slike utječe na generiranje novih slika.
Abstract (english)
This master's thesis describes the basics of convolutional neural networks. The paper describes the basics and way of working of neural networks, explains the constituent parts, the layers of the network and presents examples of the most commonly used convolutional neural networks. The purpose of the practical part is the application of a convolutional neural network in the process of neural style transfer. It is a process in which the features of one image are transferred to another image. The theoretical side and formula of the algorithm used in the neural style transfer algorithm are described. The program solution for the mentioned process was written in the Python programming language, more specifically the PyTorch library, which is used for developing and training neural networks based on deep learning models. The convolutional neural network architecture used in the solution is VGG19. The results of the solution are presented in the last part of this paper, where you can see how the generated image looks like after transferring image features from image to another and how the complexity of the image affects generating of new images.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:030103
Study programmeTitle: Graduate University Study Programme in Computer Engineering
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2022-10-18 15:02:01