Klasifikacija emocija konvolucijskim neuronskim mrežama pomoću glasovnih podataka

master's thesis

• Title (english): Convolutional Neural Networks for Emotion Recognition Tasks Using Voice/Speech Data
• Author: Nikola Vrebčević
• Mentor: Davor Petrinović (mentor)
• Committee member: Davor Petrinović (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Nikola Banić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-11, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Računala su postala neizostavan dio svakodnevice, međutim komunikacija između korisnika i računala nije prilagođena uobičajenom načinu na koji ljudi komuniciraju. Afektivno računarstvo ima cilj osposobiti računala kako bi mogla prepoznavati emocije te prilagođavati odgovor na korisničke naredbe ovisno o emocionalnom stanju korisnika. U tom procesu potrebno je odrediti emocije na neinvazivan način te se kao jedna moguća metoda pokazala detekcija emocija iz govora. U ovom radu proučava se klasifikacija emocija pomoću neuronskih mreža koje se sastoje od konvolucijskih i rekurentnih slojeva nad glasovnim podacima RECOLA baze izgovora.

Abstract (english)

In recent times computers can be found everywhere. People use them to execute certain commands but interaction between human–computer is nothing like commu- nication between people. The goal of affective computing is to develop methods and algorithms for computers in order to estimate user’s emotional state and adjust the res- ponse accordingly. One of the noninvasive paths of achieving this goal is an emotion classification from speech which is described in this master’s thesis. This was achieved by neural network consisted of convolutional and recurrent layers for classification task using RECOLA speech database.

• Keywords: glas; govor; emocije; klasifikacija emocija; konvolucijske neuronske mreže; rekurentne neuronske mreže; neuronske mreže
• Keywords (english): speech; emotions; emotion classification; convolutional neural network; recurrent neural network; neural network
• Language: croatian
• URN:NBN: urn:nbn:hr:168:869539
• Study programme: Title: Information and Communication Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-14 02:04:42