Title :
Recognizing emotions from human speech using 2-D neural classifier and influence the selection of input parameters on its accuracy
Author :
Voznak, M. ; Partila, P. ; Mehic, Miralem ; Jakovlev, S.
Author_Institution :
Fac. of Electr. Eng. & Comput. Sci., VSB-Tech. Univ. of Ostrava, Ostrava-Poruba, Czech Republic
Abstract :
This paper deals with the comparison of different methods of speech features extraction for a neural network classifier. We have used a Kohohen self-organizing feature map (SOM) for output-stage classifier which is a specific type of artificial neural nets. The result of this research deals with the accuracy of emotion classifier and compares the two input combinations.
Keywords :
emotion recognition; feature extraction; self-organising feature maps; speech recognition; 2D neural classifier; Kohohen self-organizing feature map; artificial neural nets; emotion classifier; emotion recognition; human speech; input parameter selection; neural network classifier; output-stage classifier; speech features extraction; Accuracy; Neural networks; Neurons; Speech; Speech processing; Speech recognition; Digital speech processing; emotions; fundamental frequency; neural network;
Conference_Titel :
Telecommunications Forum (TELFOR), 2013 21st
Conference_Location :
Belgrade
Print_ISBN :
978-1-4799-1419-7
DOI :
10.1109/TELFOR.2013.6716272
