Uvjetna neuronska polja za semantičku estimaciju scene monokularnom kamerom

master's thesis

• Title (english): Conditional neural fields for monocular semantic scene estimation
• Author: Josip Šestak
• Mentor: Marija Seder (mentor)
• Committee member: Ivan Petrović (član povjerenstva)
• Committee member: Luka Petrović (član povjerenstva)
• Committee member: Ivan Marković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-09-17, Croatia

Abstract

Rad se bavi primjenom uvjetnih neuronskih polja za semantičku estimaciju scene koristeći monokularnu kameru. Cilj je razviti model koji može precizno klasificirati i segmentirati različite objekte unutar jedne slike snimljene monokularnom kamerom. Uvjetna neuronska polja kombiniraju prednosti konvolucijskih neuronskih mreža i uvjetnih slučajnih polja. To omogućuje modelu da uči složene prostorne odnose između piksela u slikama.
Ključna komponenta pristupa je primjena samonadziranog učenja, metode u kojoj model koristi neoznačene podatke za generiranje vlastitih oznaka tijekom procesa obuke. Ovo smanjuje potrebu za velikim količinama ručno označenih podataka.

Abstract (english)

The work focuses on the application of conditional neural fields for semantic scene estimation using a monocular camera. The goal is to develop a model that can accurately classify and segment different objects within a single image captured by a monocular camera. Conditional neural fields combine the advantages of convolutional neural networks and conditional random fields. This allows the model to learn complex spatial relationships between pixels in images.

A key component of the approach is the application of self-supervised learning, a method in which the model uses unlabeled data to generate its own labels during the training process. This reduces the need for large amounts of manually labeled data.

• Keywords: neuronska polja; samonadzirano učenje; računalni vid
• Keywords (english): neuron fields; selfsupervised learning; computer vision
• Language: croatian
• URN:NBN: urn:nbn:hr:168:605089
• 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: 2024-10-12 22:14:43