Sinteza novih pogleda temeljena na neuronskim poljima zračenja

master's thesis

• Title (english): Novel View Synthesis Based on Neural Radiance Fields
• Author: Matej Ciglenečki
• Mentor: Marko Subašić (mentor)
• Committee member: Marko Subašić (predsjednik povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Committee member: Krešimir Križanović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-13, Croatia

Abstract

Neuronsko polje zračenja (engl. Neural Radiance Field, skraćeno NeRF) je metoda za sintetiziranje novih pogleda složene 3D scene. Rad istražuje kvalitetu sintetiziranih neviđenih pogleda i analizira mogućnosti NeRF-a za generiranje visokokvalitetnih slika iz novih pogleda. Metodom prijenosnog učenja istražena je mogućnost sintetiziranja segmentacijskih maski. Ispitano je ponašanje modela u ovisnosti o broju korištenih slika kao i mogućnost generalizacije modela na poglede koji znatnije... More odstupaju od viđenih. Provedena je analiza generalizacije modela u ovisnosti o varijabilnosti scene. Less

Abstract (english)

Neural Radiance Field (NeRF) is a method for synthesizing novel views of a complex 3D scene. The paper investigates the quality of synthesized unseen views and analyzes the capabilities of NeRF to generate high-quality images from new views. The possibility of synthesizing segmentation masks was investigated using the transfer learning method. The behavior of the model was tested depending on the number of images used, as well as the possibility of generalizing the model to views that... More differ significantly from those seen. An analysis of the generalization of the model was carried out depending on the variability of the scene. Less

• Keywords: strojno učenje; neuralna polja zračenja; sinteza novog pogleda; nerf
• Keywords (english): machine learning; neural radiance fields; novel view synthesis; nerf
• Language: croatian
• URN:NBN: urn:nbn:hr:168:921221
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 08:35:57