Samodejno prepoznavanje obraznih izrazov na iOS platformi

[diplomsko delo]

Matic Kunaver (Author), Peter Peer (Mentor)

Abstract

V tem delu je predstavljen način za samodejno prepoznavanje človeških obraznih izrazov s slik, pridobljenih s prenosnim telefonom, ki ima naložen operacijski sistem iOS. Najprej je treba določiti območje obraza. Nato je treba preoblikovati podatke v računalniku primernejšo obliko. V zadnjem koraku je treba preoblikovane podatke shraniti in izgraditi model za napovedovanje obraznih izrazov. Za napovedovanje je uporabljena metoda podpornih vektorjev, ki ima dobro razmerje med hitrostjo in natančnostjo. Končni rezultat tega dela je knjižnica, ki jo lahko vsak razvijalec priloži svojemu projektu in s pripravljenimi programskimi funkcijami dobi možnost za prepoznavo obraznih izrazov. Razvita rešitev je bila naučena in preizkušena na označenih obrazih iz baze Cohn-Kanade in iz baze tekmovanja s spletne strani Kaggle. Vsi algoritmi v tem delu se opirajo na programsko knjižnico OpenCV, razvita knjižnica pa deluje na večini prenosnih naprav družbe Apple z operacijskim sistemom iOS.

Keywords

računalniški vid;strojno učenje;obrazni izrazi;prenosni telefoni;preoblikovanje podatkov računalniški vid;preoblikovanje podatkov;računalništvo;univerzitetni študij;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2014
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[M. Kunaver]
UDC:	004.93(043.2)
COBISS:	10687316
Views:	43
Downloads:	19
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Automatic recognition of facial expressions on iOS platform
Secondary abstract:	In this diploma thesis we present an algorithm for automatic recognition of facial expressions in images from mobile devices with iOS operating system. In the first step we have to find a face region. In the next step we need to transform input data into a computer readable form. There are several methods for transforming the data. In the last step the data has to be stored and a computer model for predicting facial expressions has to be generated. Prediction is done by using Support Vector Machines. SVM method is fast and reliable. The final result of the work presented in the thesis is a library available to all developers. The library is easy to use and enables developers to get facial expression information with almost zero work. This solution was tested on labeled faces from Cohn-Kanade database and on labeled faces from Kaggle website. All algorithms in this work rely on OpenCV framework. The library works on almost all Apple mobile devices with iOS operating system.
Secondary keywords:	computer vision;machine learning;facial expressions;mobile phones;data transformation computer vision;data transformation;computer science;diploma;
File type:	application/pdf
Type (COBISS):	Undergraduate thesis
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	67 str.
ID:	24215044