prikaz prve stranice dokumenta Primjena duboke neuronske mreže u raspoznavanju objekata
  Download
PDF 1.34 MB
master's thesis
Primjena duboke neuronske mreže u raspoznavanju objekata
2019. urn:nbn:hr:200:541648
Žalac, Josip
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Cite this document

Žalac, J. (2019). Primjena duboke neuronske mreže u raspoznavanju objekata (Master's thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:541648

Žalac, Josip. "Primjena duboke neuronske mreže u raspoznavanju objekata." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2019. https://urn.nsk.hr/urn:nbn:hr:200:541648

Žalac, Josip. "Primjena duboke neuronske mreže u raspoznavanju objekata." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2019. https://urn.nsk.hr/urn:nbn:hr:200:541648

Žalac, J. (2019). 'Primjena duboke neuronske mreže u raspoznavanju objekata', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 18 April 2024, https://urn.nsk.hr/urn:nbn:hr:200:541648

Žalac J. Primjena duboke neuronske mreže u raspoznavanju objekata [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2019 [cited 2024 April 18] Available at: https://urn.nsk.hr/urn:nbn:hr:200:541648

J. Žalac, "Primjena duboke neuronske mreže u raspoznavanju objekata", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2019. Available at: https://urn.nsk.hr/urn:nbn:hr:200:541648

Please login to the repository to save this object to your list.
Metadata
TitlePrimjena duboke neuronske mreže u raspoznavanju objekata
Title (english)Implementation of deep neural networks in object recognition
AuthorJosip Žalac
MentorEmmanuel Karlo Nyarko (mentor)
Committee memberEmmanuel Karlo Nyarko (predsjednik povjerenstva)
Committee memberDamir Filko (član povjerenstva)
Committee memberPetra Đurović (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Computer Engineering and Automation)
(Chair of Automation and Robotics)
Osijek
Defense date and country2019-10-15, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
U ovom radu opisane su arhitekture i način rada konvolucijskih neuronskih mreža te je detaljno objašnjena zadaća svakog sloja unutar konvolucijske neuronske mreže. Također, opisana je metoda transfer učenja i načini upotrebljavanja u praksi. Detaljno je prikazana arhitektura i način rada modela upotrebljavanih za transfer učenje, te njihova implementacija i testiranje na skupu podataka dviju klasa. Obavljeno je učenje vlastitiog modela te na temelju rezultata, najbolji model se upotrebljavao za daljnje podešavanje da se dobije što veća preciznost predikcije. Model je također testiran na skupu podataka te su rezultati uspoređeni
Abstract (english)
In this master thesis, an architecture and the way convolutional neural networks work are described in detail. Also, each layer of Convolutional neuaal network and it's task within network is shown. Aside from that, method of Transfer learning is introduced, and it's usage in practice. Description of one of the covered methods and it's implementation is shown on datasets used throughout this thesis, containing two classes. Also, training of our own models is shown with variety of architectures using the same dataset. Model with best performanse was later used to fine-tune and to try different optimizing algorithms. After fine-tuning, results were compared to pre-trained models
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:541648
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 on2019-11-14 12:05:10