Speech emotion recognition using RBF kernel of LIBSVM

Author

Chavhan, Y.D. ; Yelure, B.S. ; Tayade, K.N.

Author_Institution

GCE, Karad, Karad, India

fYear

2015

fDate

26-27 Feb. 2015

Firstpage

1132

Lastpage

1135

Abstract

Automatic Speech Emotion Recognition (SER) is a current research topic in the field of Human Computer Interaction (HCI) with wide range of applications. The speech features such as, Mel Frequency cepstrum coefficients (MFCC) and Mel Energy Spectrum Dynamic Coefficients (MEDC) are extracted from speech utterance. The LIBSVM is used as classifier to identify different emotional states such as anger, happiness, sadness, neutral, fear, from Berlin emotional database. The results are taken by using RBF kernel of LIBSVM. It gives 93.75% recognition accuracy for RBF kernel.

Keywords

cepstral analysis; emotion recognition; feature extraction; human computer interaction; radial basis function networks; regression analysis; signal classification; speech recognition; support vector machines; Berlin emotional database; HCI; LIBSVM; MEDC; MFCC; RBF kernel; SER; anger; automatic speech emotion recognition; emotional states; fear; happiness; human computer interaction; mel energy spectrum dynamic coefficients; mel frequency cepstrum coefficients; neutral; sadness; speech feature extraction; speech utterance; Emotion recognition; Feature extraction; Kernel; Mel frequency cepstral coefficient; Speech; Speech recognition; Support vector machines; Emotion Recognition; LIBSVM; MFCC and MEDC; RBF; Speech emotion;

fLanguage

English

Publisher

ieee

Conference_Titel

Electronics and Communication Systems (ICECS), 2015 2nd International Conference on

Conference_Location

Coimbatore

Print_ISBN

978-1-4799-7224-1

Type

conf

DOI

10.1109/ECS.2015.7124760

Filename

7124760