Analiza sorodstva na podlagi slik uhljev

magistrsko delo

Grega Dvoršak (Author), Vitomir Štruc (Mentor), Žiga Emeršič (Co-mentor)

Abstract

Področje biometrije je v zadnjem desetletju močno napredovalo, še vedno pa lahko najdemo prostor za izboljšave. Ena od potencialno uporabnih modalnosti so uhlji, ki jih že lahko uporabljamo v različnih aplikacijah na področju varnosti in nadzora v kombinaciji z drugimi modalnostmi. Uhlji imajo značilnosti s katerimi je mogoče razmeroma dobro razlikovati med ljudmi, zato jih lahko uporabimo za razpoznavanje identitete oseb. V magistrskem delu nas zanima, če je mogoče slike uhljev uporabiti za razpoznavanje sorodstva na podlagi njihovih značilnosti. Poleg analize razpoznavanja sorodstva je eden od prispevkov tudi podatkovna baza, ki vsebuje 19 družin s slikami uhljev za vsakega družinskega člana. Slike v podatkovni bazi so ročno anotirane z očrtanimi pravokotniki, ki služijo poravnavi uhljev, oblikovan pa je tudi seznam slik slabe kvalitete, s katerim lahko take slike izločimo iz učenja. Za razpoznavanje sorodstva je oblikovan osnovni model siamske nevronske mreže, ki uporablja 5 različnih hrbtenic, s katerimi opravimo različne eksperimente. Rezultati kažejo, da so uhlji primerna modalnost za razpoznavanje sorodstva, saj 4 od 5 napovednih modelov doseže uspešnost nad 60% površine pod krivuljo ROC, najboljša uspešnost pa je 74,1%.

Keywords

biometrija uhljev;sorodstvo;računalništvo in informatika;magisteriji;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.09 - Master's Thesis
Organization:	UL FE - Faculty of Electrical Engineering
Publisher:	[G. Dvoršak]
UDC:	004.93:57.087.1(043.2)
COBISS:	89826819
Views:	244
Downloads:	27
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Kinship Analysis Using Ear Images
Secondary abstract:	The area of biometrics has seen large advancement in the last decade, although there still remains room for improvement. One of the potentially very useful modalities are ears, which can be used in various different applications in the security and surveillance fields used in combination with other modalities. Ears possess characteristics which can be used to determine the identity of a person relatively well. Therefore, ears can be used in identity recognition tasks. The topic of this master's thesis is to determine whether the characteristics of ears in ear images can be used for kinship analysis. Along with the analysis, a contribution to the thesis is also a dataset consisting of 19 families with ear images for each family member. The images also have manually annotated bounding boxes, which are used for the alignment of the ears, and additionally we provide a list of bad quality images, which can be used to exclude the images from the learning process. For the kinship analysis, a Siamese neural network model is developed, for which 5 different backbones can be used to perform kinship verification via various experiments. Results show that ears are a suitable modality for kinship verification as 4 out of 5 prediction models reach a performance of over 60% of area under the ROC curve, the best performance being 74.1%.
Secondary keywords:	biometrics;kinship;ear biometrics;computer vision;computer science;computer and information science;master's degree;Biometrija;Biometrična identifikacija;Računalniški vid;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS):	Master's thesis/paper
Study programme:	1000471
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	58 str.
ID:	14083830