The Context Sets the Tone: A Literature Review on Emotion Recognition from Speech Using AI

Customers’ emotions play a crucial role in the service industry. The better the staff understands the customer, the better their service. Human emotions elicit measurable speech markers, such as increased speech rate or higher pitch, and AI can interpret these signals. In recent years, significant progress has been made in automatically recognizing basic emotions such as joy, anger, etc. However, there is a lot of disagreement regarding evaluating the crucial feature types and the feature dimensions to be analyzed. By utilizing a systematic literature review of 81 articles, this article analyses the specification of context and DSR implications for feature types and emotion dimensions for emotional speech analysis. The results show that these are generally insufficiently specified. Accordingly, this paper aims to optimize the potential for generalization of DSR results and thereby improves theory building in this discipline.