diplomsko delo
Tadej Jerovšek (Author), Božidar Potočnik (Mentor)

Abstract

V tem diplomskem delu se ukvarjamo z detektiranjem oči v barvnih digitalnih posnetkih ter segmentacijo le-teh, saj večina obstoječih metod vrača kot rezultat le oklepajočo škatlo očesa. Pri reševanju tega problema smo za detekcijo beločnice uporabili barvni prostor HSV, za iskanje šarenice smo uporabili prileganje modela, pri določitvi zenice pa smo uporabili nevronsko mrežo. Naš algoritem smo nato validirali na bazi 50 slik, kjer smo ugotavljali uspešnost za posamezne komponente očesa. Pri tem smo ugotovili, da pri večini razpoznamo komponento v dobri meri, vendar pri nekaterih razpoznamo tudi napačni del slike. Kot izhod oblikujemo maske posameznih komponent, ki jih lahko uporabimo za nadaljnja dela z očmi.

Keywords

računalniški vid;prepoznavanje oči;razpoznavanje vzorcev;digitalna obdelava slik;Houghova transformacija;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [T. Jerovšek]
UDC: 004.932.72(043.2)
COBISS: 17991958 Link will open in a new window
Views: 1688
Downloads: 91
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: AUTOMATED EYE RECOGNITION FROM DIGITAL IMAGES
Secondary abstract: In the diploma thesis we deal with detecting eyes in color digital images and their segmentation, because most existing methods return only a bounding box as a result. To solve this problem, we used the HSV color space for detecting the sclera, model fitting for finding the iris and we used a neural network for determining the pupil. We then validated our algorithm on a base of 50 images, where we determined its success rate for individual components of the eye. With this we determined that in most pictures we identify the component in a good measure, but in some we also identify a false part of the image. We form masks of individual components as an output, which can then be used for further work with eyes.
Secondary keywords: computer vision;eye recognition;pattern recognition;digital image processing;Hough transform;
URN: URN:SI:UM:
Type (COBISS): Bachelor thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko
Pages: X, 41 str.
ID: 8729414