A Priori LSTM based-Emotion recognition in speech signals
Author Name : Puneet Singh Lamba, Charanpreet Kaur
Voice has been considered as the most natural medium for communicating and expressing emotions. As a result, an ample amount of effort has been put in to make machines function on voice-based commands. Emotion recognition in speech plays a significant role in improving human-machine communication. In this paper, we have presented the emotion recognition performance of various classifiers on the eNTERFACE database. The emotions differentiated by classifiers are mainly calm, happy, sad, angry, surprise, etc. Features such as pitch, energy, and Mel Frequency Cepstrum Coefficients (MFCC) of these speech samples have been considered in our research work.
Keywords: Emotion recognition, MFCC features, Convolutional neural network, LSTM network.