Prepoznava oseb na odprtih množicah slik uhljev

diplomsko delo

Sebastijan Trojer (Author), Žiga Emeršič (Mentor), Blaž Meden (Co-mentor)

Abstract

Diplomska naloga se ukvarja z izzivom prepoznavanja oseb v okolju odprtih množic podatkov. V nalogi je predstavljen pristop, ki temelji na uporabi metod za prepoznavo na odprtih množicah z uporabo siamskih nevronskih mrež in trojne cenilne funkcije. Glavni cilj je bil optimizirati obstoječe modele prepoznavanja s pomočjo naprednih tehnik in algoritmov ter primerjati rezultate klasičnih in optimiziranih modelov. Rezultati kažejo, da optimizirani modeli ne presežejo zmogljivosti klasičnih, kar kaže na težavo z uporabo optimiziranih pristopov pri reševanju problema prepoznavanja oseb na odprtih množicah podatkov, omogočajo pa bolj časovno učinkovito implementacijo. Pridobljeno znanje lahko služi kot temelj za nadaljnje raziskave in razvoj na tem področju.

Keywords

prepoznavanje oseb;nevronske mreže;uhlji;slike uhljev;univerzitetni študij;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2024
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[S. Trojer]
UDC:	004.93:57.087.1(043.2)
COBISS:	208484355
Views:	119
Downloads:	36
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Openset person recognition on ear images
Secondary abstract:	The thesis addresses the challenge of recognizing individuals in an open set data environment. It presents an approach based on the use of methods for open set recognition using siamese neural networks and triplet loss. The main objective was to optimize existing recognition models using advanced techniques and algorithms and to conduct a comparative analysis between classical and optimized models. The results show that the optimized models do not exceed performance of classical models, indicating the proposed methods do not have the potential to be used in solving the problem of recognizing individuals in open set data environments, however they offer an implementation with better time performance. The acquired knowledge can serve as a foundation for further research and development in this field.
Secondary keywords:	deep learning;biometry;computer vision;computer and information science;diploma;Biometrična identifikacija;Računalniški vid;Globoko učenje (strojno učenje);Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS):	Bachelor thesis/paper
Study programme:	1000468
Embargo end date (OpenAIRE):	1970-01-01
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	1 spletni vir (1 datoteka PDF (39 str.))
ID:	24862809