A Review - ML and DL Classifiers for Emotion Detection in Audio and Speech Data

Emotion is an inherited characteristic in all modes of communication and is a natural part of human behavior. The human brain, can sense others' expressions, feelings, and moods due to their brain knowledge, making us more reasonable and understanding. Machines, on the other hand, have a hard time understanding content-based information like text, music, or video. It is a requirement of the period that machines be educated to accurately interpret emotions in order to improve understanding and minimize misunderstanding. Furthermore, audio emotion analysis has several applications in a variety of industries, such healthcare, banking, defense, and information technology. Text emotions, on the other hand, are easier to comprehend because tone and pitch are irrelevant, but in audio emotion analysis, both aspects must be considered for higher accuracy. There are also other elements that degrade accuracy, such as noise, disruption, and various gaps in transmission. Making a machine recognize the respondent's emotion is a difficult task. This review investigates the use of ML and DL algorithms to recognize emotion in speech data.