Uklanjanje neočekivanih artefakata s fotografija ljudskih lica korištenjem dubokog učenja

undergraduate thesis

• Title (english): Removing unexpected artifacts from photographs of human faces using deep learning
• Author: Karlo Marković
• Mentor: Domagoj Vlah (mentor)
• Committee member: Domagoj Vlah (predsjednik povjerenstva)
• Committee member: Stjepan Šebek (član povjerenstva)
• Committee member: Iva Manojlović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-08-31, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj, a ujedno i motivacija ovog rada je uklanjanje neočekivanih artefakata sa slika ljudskih lica. Rješenju navedenog problema pristupa se korištenjem dubokog učenja. Preciznije, rad prikazuje koliko dobro arhitektura neuronske mreže, varijacijski autoenkoder, vrši rekonstrukciju oštećene slike. Model je implementiran uz pomoć biblioteke za duboko učenje Pytorch, a treniran je na dopunjenom skupu podataka CelebA. U radu se također navode opći podatci o dubokom učenju, arhitekturi autoenkoder, biblioteci Pytorch i skupu podataka CelebA.

Abstract (english)

The goal, and at the same time the motivation of this work, is to remove unexpected artifacts from images of human faces. It is proposed to get to the solution using deep learning. More precisely, the paper shows how well the architecture of the neural network, the variational autoencoder, reconstructs a damaged image. The model was implemented using the deep learning library Pytorch, and was trained on the extended CelebA dataset. The paper also provides general information about deep learning, the autoencoder architecture, the Pytorch library, and the CelebA dataset.

• Keywords: duboko učenje; neuronske mreže; varijacijski autoenkoder; biblioteka Pytorch; skup podataka CelebA
• Keywords (english): deep learning; neural networks; variational autencoder; Pytorch library; CelebA dataset
• Language: croatian
• URN:NBN: urn:nbn:hr:168:419976
• 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: 2023-10-11 09:42:11