From the point of view of practical value, the definition of emotions in a person's voice can be applied in various areas related to both the transmission of audio messages and online communication: such areas include medicine, security, economics, education, etc. As a striking example, we can provide an assessment of the quality of work of call-center operators, as well as the services / goods they offer. So the presence of signals that the client is experiencing negative emotions, for example, anger, can indicate possible problems with the operators. In this paper, a neural network approach will be considered for automatically determining a person's emotions from his speech.

Keywords: neural network, emotion detection, speech, classification, deep learning, convolutional model