Research Article
Acoustic Intelligence In Conversational Solutions Emotion Detection From Speech
@INPROCEEDINGS{10.4108/eai.7-12-2021.2314555, author={A Chitra and Ashok Raj and Vishnubalaji R K}, title={Acoustic Intelligence In Conversational Solutions Emotion Detection From Speech}, proceedings={Proceedings of the First International Conference on Combinatorial and Optimization, ICCAP 2021, December 7-8 2021, Chennai, India}, publisher={EAI}, proceedings_a={ICCAP}, year={2021}, month={12}, keywords={emotion recognition; iemocap; recurrent neural networks; long short term memory}, doi={10.4108/eai.7-12-2021.2314555} }- A Chitra
Ashok Raj
Vishnubalaji R K
Year: 2021
Acoustic Intelligence In Conversational Solutions Emotion Detection From Speech
ICCAP
EAI
DOI: 10.4108/eai.7-12-2021.2314555
Abstract
Speech is one of the natural ways of expressing ourselves as humans. It is defined as the ability to convey thoughts, ideas, or other information by means of articulating sound into meaningful words. In computer technology, speech processing paved the way for an infinite number of innovations such as Siri,Alexa, etc., which are artificial intelligence systems that are embedded in mobile devices that recognize and respond to human commands. Speech emotion can be perceived with the message of utterance and independent language of utterance. Speech emotion recognition systems are the collection of methodologies that process and classify speech signals to detect human emotions. Speech emotion recognition systems use speech signals labeled with emotions as a dataset. In this proposed system, supervised deep learning techniques such as Long Short Term Memory (LSTM) and Bidirectional LSTM models are used as classifiers to classify emotions. The emotions that are considered for classification are angry, neutral, sad, and excited. Speech emotion recognition has its applications in robotics, call centers, etc