Speech Emotion Recognition by using SVM-Classifier
Vaishali M. Chavan1, V. V. Gohokar2
1Vaishali M. Chavan, Electronics & Telecommunication Engg., Amravati University/ S. S. G. M. C. E, Shegaon.
2V. V. Gohokar, Electronics & Telecommunication Engg., Amravati University/ S. S. G. M. C. E, Shegaon.
Manuscript received on may 25, 2012. | Revised Manuscript received on June 22, 2012. | Manuscript published on June 30, 2012. | PP: 11-15 | Volume-1 Issue-5, June 2012 | Retrieval Number: E0367041512/2012©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Automatic emotion recognition in speech is a current research area with a wide range of applications in human-machine interactions. This paper uses the support vector machine (SVM), to classify five emotional states: anger, happiness, sadness, surprise and a neutral state. The classification performance of the selected feature subset was done with that of the Mel frequency cepstrum coefficients (MFCC), Periodicity Histogram and Fluctuation Pattern. Within the method based on SVM, a new method by using Multi-class SVM is used as a classifier. Experiments were conducted on the Danish Emotion Speech (DES) Database. The recognition rates by using SVM classifier were 68 %, 60 %, 55.40 % and 60 % for Linear, Polynomial, RBF, and Sigmoid Kernel Function respectively. The recognition rates by Multiclass SVM using Linear, Polynomial, RBF and Sigmoid kernel function for Danish database for Periodicity Histogram are 64.77 %, 78.41 %, 79.55 % and 78.41% respectively.
Keywords: Emotion recognition, Mel frequency cepstrum coefficients (MFCC), Support Vector Machine.