Sustav za obradu slike kojim se postiže prijenos stila

undergraduate thesis

• Title (english): Image Processing System for the Style Transfer
• Author: Filip Barbarić
• Mentor: Mislav Grgić (mentor)
• Committee member: Mislav Grgić (predsjednik povjerenstva)
• Committee member: Sonja Grgić (član povjerenstva)
• Committee member: Tomislav Kos (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Prijenos stila je obrada slike s kojom se sadržaj neke slike stilizira zadanim stilom. Takve metode danas se najčešće koriste u računalnoj grafici i kao pomoć pri animiranju. Postoji mnogo metoda s kojim se može prenijeti no neuroprijenos stila danas daje najbolje rezultate i najslobodnije rezultate. Prvi prijenos stila pomoću neuronske mreže ostvario je Leon Gatys. Sadržaj slike lako se izračunavao no problem je bio stil. Gatys je predstavio metodu računanja stila preko gramovih matrica. Njegova metoda prijenosa stila provodi se tako da za definiranu funkciju gubitka optimiziramo našu generiranu sliku. Ta funkcija gubitka definirana je sa stilom i sadržajem. Ubrzo su se razvile i druge metode koje rade brže.

Abstract (english)

Style transfer is the processing of an image with which the content of an image is styled by a given style. Such methods are most commonly used today in computer graphics and as an aid to animation. There are many methods by which it can be transmitted but neurotransmission of style today gives the best results and the freest results. The first transmission of style using a neural network was achieved by Leon Gatys. The content of the image was easy to calculate but the problem was the style. Gatys introduced a method of calculating style over gram matrices. Its style transfer method is implemented by optimizing our generated image for a defined loss function. This loss function is defined with style and content. Other methods that work faster soon developed.

• Keywords: Neuronska mreža; Duboko učenje; Konvolucija; Konvolucijske neuronske mreže; VGG19; Neuronski prijenos stila
• Keywords (english): Neural Network; Deep Learning; Convolution; Convolutional Neural Networks; VGG19; Neural Style Transfer
• Language: croatian
• URN:NBN: urn:nbn:hr:168:489164
• 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: 2023-01-14 20:03:18