prikaz prve stranice dokumenta Oblici i njihova semantika u digitalnim sadržajima
No public access
master's thesis
Oblici i njihova semantika u digitalnim sadržajima
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2023. urn:nbn:hr:168:443029
Nazlić, Dino
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Nazlić, D. (2023). Oblici i njihova semantika u digitalnim sadržajima (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:443029

Nazlić, Dino. "Oblici i njihova semantika u digitalnim sadržajima." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2023. https://urn.nsk.hr/urn:nbn:hr:168:443029

Nazlić, Dino. "Oblici i njihova semantika u digitalnim sadržajima." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2023. https://urn.nsk.hr/urn:nbn:hr:168:443029

Nazlić, D. (2023). 'Oblici i njihova semantika u digitalnim sadržajima', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 24 June 2024, https://urn.nsk.hr/urn:nbn:hr:168:443029

Nazlić D. Oblici i njihova semantika u digitalnim sadržajima [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2023 [cited 2024 June 24] Available at: https://urn.nsk.hr/urn:nbn:hr:168:443029

D. Nazlić, "Oblici i njihova semantika u digitalnim sadržajima", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:168:443029

Metadata
TitleOblici i njihova semantika u digitalnim sadržajima
Title (english)Forms and their semantics in digital contents
AuthorDino Nazlić
MentorDragan Jevtić (mentor)
Committee memberDragan Jevtić (predsjednik povjerenstva)
Committee memberKrešimir Pripužić (član povjerenstva)
Committee memberŽeljka Car (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2023-07-04, Croatia
Abstract
Ovaj rad opisuje osnovne koncepte klasifikacije slika korištenjem konvolucijskih neuronskih mreža kao i postupak izrade algoritma za klasifikaciju prometnih znakova. Za izradu algoritma korišten je transformirani GTSRB skup podataka sa oko 40 000 slika. Preciznost i iznos funkcije pogreške ispitani su na modelima različitih struktura te je izvršen i odabir hiperparametara kako bi se odredio model koji daje najveću preciznost i najmanju vrijednost funkcije pogreške. Izdvojeni model sastoji se od 3 konvolucijska i 2 potpuno povezana sloja, slojeva sažimanja maksimumom i regularizacije isključivanjem čvorova. Model postiže preciznost od oko 96 % na neviđenom skupu to jest skupu za validaciju.
Abstract (english)
This master's thesis outlines the fundamental ideas of convolutional neural network image classification as well as the detailed procedure for developing a CNN traffic sign recognizer. About 40 000 images taken from the preprocessed GTSRB dataset were used for developing the model. To determine accuracy and loss values as well as the optimal combination of hyperparameters, models with different structures were tested. The chosen model consists of 3 convolutional and 2 fully conected layers, maximum pooling layers, and regularization through the use of dropout. The model's accuracy on the validation set was 96%.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:443029
Study programmeTitle: Computing
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 conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2023-10-11 09:54:45