Speech Emotion Recognition Systems: A Comprehensive Review on Different Methodologies

As humans, speech is the common as well as a natural way of expressing ourselves. Speech Emotion Recognition (SER) systems can be defined as an assortment of methods processes and classifies speech signals for the detection of associated emotions. Automatic emotion recognition is the technique of identification of human emotions from various signals like speech, facial expression and text. Collection of such signals and labelling them is often tiresome and needs proficient knowledge. This paper deals with the different types of open source speech emotion datasets of various languages and recent literature survey in the area of speech emotion recognition that employs a number of machine learning approaches with an objective of enhancing the classification accuracy. The paper prudently aims at identifying and synthesizing contemporary pertinent literature associated to the SER systems with different methodologies or design components, thus providing the researchers with an up-to-date understanding of the research topic in the field of SER.