Primjena umjetnih neuronskih mreža za prepoznavanje odabranih riječi hrvatskog govora

master's thesis

• Title (english): Applying Artificial Neural Networks for Recognizing Selected Words in Croatian
• Author: Andrea Zlatić
• Mentor: Marin Vuković (mentor)
• Committee member: Marin Vuković (predsjednik povjerenstva)
• Committee member: Dragan Jevtić (član povjerenstva)
• Committee member: Krešimir Pripužić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-09, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj rada bio je istražiti postojeće primjene umjetnih neuronskih mreža u
raspoznavanju govora. U radu je dan pregled nekih sličnih radova te je
implementiran sustav za prepoznavanje govornih naredbi u hrvatskom jeziku
pomoću programskog okvira Tensorflow. Pokazalo se da je moguće prepoznati
govorne naredbe prikazane spektrogramom pomoću konvolucijske neuronske
mreže. Moguća poboljšanja su proširenje skupa korištenih riječi te produbljenje
korištene arhitekture, kao i testiranje drugačijeg prikaz podataka.

Abstract (english)

The goal of this dissertation was to research existing uses of artificial neural
networks in speech recognition systems. An overview of similar works was given
as well as an implementation of a system for recognizing speech commands in
the Croatian language using the Tensorflow framework. It is shown that it is
possible to recognize speech commands in form of a spectrogram by using a
convolutional neural network. Possible improvements are expanding the set of
used words and increasing the depth of the used network architecture, as well as
testing a different input representation.

• Keywords: Umjetne neuronske mreže; konvolucijske neuronske mreže; prepoznavanje govora; spektrogram
• Keywords (english): Artificial neural networks; convolutional neural networks; speech recognition; spectrogram
• Language: croatian
• URN:NBN: urn:nbn:hr:168:229733
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-14 02:19:40