ABSTRACT
Emotion recognition is the process to detect, evaluate, interpret, and respond to people's emotional states and emotions, ranging from happiness to fear to humiliation. The COVID- 19 epidemic has provided new and essential impetus for emotion recognition research. The numerous feelings and thoughts shared and posted on social networking sites throughout the COVID-19 outbreak mirrored the general public's mental health. To better comprehend the existing ecology of applied emotion recognition, this work presents an overview of different emotion acquisition tools that are readily available and provide high recognition accuracy. It also compares the most widely used emotion recognition datasets. Finally, it discusses various machine and deep learning classifiers that can be employed to acquire high level features for classification. Different data fusion methods are also explained in detail highlighting their benefits and limitations. © 2022 The Author(s)