Academic Journals Database
Disseminating quality controlled scientific knowledge

Speech based Emotion Recognition with Gaussian Mixture Model

Author(s): Nitin Thapliyal , Gargi Amoli

Journal: International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
ISSN 2278-1323

Volume: 1;
Issue: 5;
Start page: 065;
Date: 2012;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Speech | Gaussian Mixture Model | Vocal | Emotion Recognition | Linear predictive.

This paper is mainly concerned with speech basedemotion recognition. The main work is concerned withGaussian mixture model (GMM model) which allows trainingthe desired data set from the databases. GMM are known tocapture distribution of data point from the input feature space,therefore GMM are suitable for developing emotion recognitionmodel when large number of feature vector is available. Given aset of inputs, GMM refines the weights of each distributionthrough expectation-maximization algorithm. Once a model isgenerated, conditional probabilities can be computed for testpatterns (unknown data points). Expectation maximization(EM) algorithm is used for finding maximum likelihoodestimates of parameters in probabilistic models. MoreoverLinear Predictive (LP) analysis method has been chosen forextracting the emotional features because it is one of the mostpowerful speech analysis techniques for estimating the basicspeech analysis techniques for estimating the basic speechparameter such as pitch, formants, spectra, vocal tractfunctions and for representing speech by low bit ratetransmission for storage. Speakers are made to involve inemotional conversation with the anchor, where differentcontextual situations are created by the anchor through theconversation to elicit different emotions from the subject,without his/her knowledge.
Save time & money - Smart Internet Solutions      Why do you need a reservation system?