Classification Methods Accuracy for Speech Emotion Recognition System

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989100%3A27240%2F14%3A86090869" target="_blank" >RIV/61989100:27240/14:86090869 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1007/978-3-319-07401-6_44" target="_blank" >http://dx.doi.org/10.1007/978-3-319-07401-6_44</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-319-07401-6_44" target="_blank" >10.1007/978-3-319-07401-6_44</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Classification Methods Accuracy for Speech Emotion Recognition System

Popis výsledku v původním jazyce

Emotional state classification of human speech and recognition accuracy of the classifiers is disclosed in this paper. Recent developments in speech recognition places more emphasis on the extraction of information about the speech source. This means obtain information about who and how it was said. This article describes research which seeks to recognize the information from speaking, emotional state in particular. Emotional state is recognized by using different classifiers and features of speech by nowadays known systems. Berlin database of emotional recordings was used to train and test the system. Mel-frequency spectral coefficients and dynamic coefficients were extracted from the audio signal of the database. For classification were used Gaussian Mixture Model, k-Nearest Neighbours and Artificial Neural Networks methods. The main effort of this research is to examine the accuracy and usability of classifying methods for detection of human stress status from his speech.

Název v anglickém jazyce

Classification Methods Accuracy for Speech Emotion Recognition System

Popis výsledku anglicky

Emotional state classification of human speech and recognition accuracy of the classifiers is disclosed in this paper. Recent developments in speech recognition places more emphasis on the extraction of information about the speech source. This means obtain information about who and how it was said. This article describes research which seeks to recognize the information from speaking, emotional state in particular. Emotional state is recognized by using different classifiers and features of speech by nowadays known systems. Berlin database of emotional recordings was used to train and test the system. Mel-frequency spectral coefficients and dynamic coefficients were extracted from the audio signal of the database. For classification were used Gaussian Mixture Model, k-Nearest Neighbours and Artificial Neural Networks methods. The main effort of this research is to examine the accuracy and usability of classifying methods for detection of human stress status from his speech.

Druh

D - Stať ve sborníku

CEP obor

JC - Počítačový hardware a software

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2014

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Nostradamus 2014: prediction, modeling and analysis of complex systems

ISBN

978-3-319-07400-9

ISSN

2194-5357

e-ISSN

—

Počet stran výsledku

9

Strana od-do

439-447

Název nakladatele

Springer

Místo vydání

Heidelberg

Místo konání akce

Ostrava

Datum konání akce

23. 6. 2014

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—