Modelling of emotion recognition system from speech using MFCC features

• Main Authors: Bhimavarapu, John Philip, Sarvana, Kalyan, Achanta, Vamsi Krishna Sai, Kadiyala, Chaitanya, Yadhavkareti, Chakradhar
• Format: Conference Proceeding
• Language: English
• Subjects: Digital signal processing; Emotion recognition; Emotions; Feature extraction; Man-machine interfaces; Speech; Speech recognition; Voice communication
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0066503

Speech is an advanced signal consisting of varied data, regarding the message to be communicated, speaker, language, region, emotions etc. Speech process is one among the vital branches of digital signal processing and finds applications in Human Machine interface, Telecommunication, Audio mining, Security etc., Speech recognition is vital for natural interaction between human and machine. In speech emotion recognition, the emotion state of a speaker is extracted from his or her speech. The acoustic characteristic of the speech signal is Feature. Feature extraction is the method that extracts a little quantity of information from the speech signal that may later be used to represent speaker. Several feature extraction strategies are implemented as of now and Mel Frequency Cepstral coefficient (MFCC). This paper presents speaker emotions recognized by using the information extracted from the speaker speech signal. Mel Frequency Cepstral coefficient (MFCC) technique is employed to acknowledge feeling of a speaker from their voice. The designed system was implemented for Happy, sad and anger emotions and the potency was found to be about 82% for sad, 74% for angry, and 72% for happy.