International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 42 - Issue 10 |
Published: March 2012 |
Authors: N. Murali Krishna, Y. Srinivas, P. V. Lakshmi |
![]() |
N. Murali Krishna, Y. Srinivas, P. V. Lakshmi . An Emotion Recognition System based on Right Truncated Gaussian Mixture Model. International Journal of Computer Applications. 42, 10 (March 2012), 41-45. DOI=10.5120/5733-7816
@article{ 10.5120/5733-7816, author = { N. Murali Krishna,Y. Srinivas,P. V. Lakshmi }, title = { An Emotion Recognition System based on Right Truncated Gaussian Mixture Model }, journal = { International Journal of Computer Applications }, year = { 2012 }, volume = { 42 }, number = { 10 }, pages = { 41-45 }, doi = { 10.5120/5733-7816 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2012 %A N. Murali Krishna %A Y. Srinivas %A P. V. Lakshmi %T An Emotion Recognition System based on Right Truncated Gaussian Mixture Model%T %J International Journal of Computer Applications %V 42 %N 10 %P 41-45 %R 10.5120/5733-7816 %I Foundation of Computer Science (FCS), NY, USA
This article address a novel emotion recognition system based on the Truncated Gaussian mixture model . The proposed system has been experimented over an gender independent emotion recognition database In the recent past, many models have been listed in the literature based on the emotion recognition, but these papers are more focused towards the speech, ignoring the emotion of the speaker at the time of speech which may be of significant importance at some particular instances such as BPO. To overcome this, we present a model using Right Truncated Gaussian mixture model and K-means algorithm to classify the emotion speeches. MFCC features of the emotions are extracted. The proposed system has been experimented over an gender independent emotion recognition database. The results obtained are evaluated using a confusion Matrix and compared with that of the Gaussian mixture model . Our model achieved a recognized rate of above 90 %.