prikaz prve stranice dokumenta Korekcija boja u slici dubokim neuronskim mrežama
No public access
undergraduate thesis
Korekcija boja u slici dubokim neuronskim mrežama
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2018. urn:nbn:hr:168:030270
Bago, Marko
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Bago, M. (2018). Korekcija boja u slici dubokim neuronskim mrežama (Undergraduate thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:030270

Bago, Marko. "Korekcija boja u slici dubokim neuronskim mrežama." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2018. https://urn.nsk.hr/urn:nbn:hr:168:030270

Bago, Marko. "Korekcija boja u slici dubokim neuronskim mrežama." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2018. https://urn.nsk.hr/urn:nbn:hr:168:030270

Bago, M. (2018). 'Korekcija boja u slici dubokim neuronskim mrežama', Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 25 November 2024, https://urn.nsk.hr/urn:nbn:hr:168:030270

Bago M. Korekcija boja u slici dubokim neuronskim mrežama [Undergraduate thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2018 [cited 2024 November 25] Available at: https://urn.nsk.hr/urn:nbn:hr:168:030270

M. Bago, "Korekcija boja u slici dubokim neuronskim mrežama", Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2018. Available at: https://urn.nsk.hr/urn:nbn:hr:168:030270

Metadata
TitleKorekcija boja u slici dubokim neuronskim mrežama
Title (english)Color Image Correction Using Deep Neural Networks
AuthorMarko Bago
MentorSven Lončarić (mentor)
Committee memberSven Lončarić (predsjednik povjerenstva)
Committee memberNikola Banić (član povjerenstva)
Committee memberJelena Ivanković (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2018-07-13, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
U ovome radu proučavamo korištenje duboke neuronske mreže za estimiranje iluminacije scene. Koristimo mrežu s jednim konvolucijskim slojem sa udruživanjem s maksimalnom vrijednošću, jedan potpuno povezani sloj te izlaz s tri čvora (jedan za svaku vrijednost kanala boje). Za treniranje mreže koristimo slike iz baze koja sadrži istinite vrijednosti iluminacije scena svih slika. Mreža na ulaz prima uzorke slika te daje lokalna predviđanja osvjetljenja za svaki uzorak. Globalno predviđanje osvjetljenja dobiva se jednostavno uzimanjem srednje vrijednosti svih lokalnih predviđanja. Predviđenu boju osvjetljenja scene tada možemo iskoristiti za uklanjanje neželjenog utjecaja boje osvjetljenja kromatskom adaptacijom. Korištenje dubokih mreža za rješavanje problema estimacije osvjetljenja scene daje izvrsne rezultate u usporedbi s ostalim metodama.
Abstract (english)
In this work we describe using a deep neural network for estimating the scene illumination. The network we use consists of one convolutional layer with max pooling, one fully connected layer and an output layer with three nodes (one for each color channel). For training the network we use an image dataset that contains groundtruth scene illumination values for each image. The network takes image patches as input and gives local illumination predictions for every patch. The global estimate is calculated simply by taking the average of all the local estimates. The predicted scene illumination color values can then be used to remove unwanted illumination effects using chromatic adaptation. Using deep neural networks for solving the problem of estimating the scene illumination produces excellent results compared to other methods.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:030270
Study programmeTitle: 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 resourceText
File originBorn digital
Access conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2019-03-13 18:45:09