AI methods for emotional contagion recognition

Emotional contagion represents a social phenomenon that occurs when a connection is established between two individuals, it is a slight manifestation that consists in the occurrence of the same shared emotion in the context of relationships. In the last years it has been studied and introduced in many research papers: the nearest declination of this is represented by applications and technologies of emotion recognition, trough that it is possible to study the phenomenon in a scientific and less subjective way producing experimental datasets. In this scenario methods and applications of Artificial Intelligence could help to handle this amount of data and their classification tasks may give a good help for the researchers and experiments on the deep comprehension and several ways of manifestations of emotional bonds among individuals. This paper tries to present the most diffused AI-based methods used for emotion recognition scopes, which is a wide field of studies, but applicated to social emotions (specifically yawn and laughter). Firstly, a theory introduction on emotions topics and emotional contagion is presented in a general way to present the focus of the research, followed by a general analysis on characteristics about context and sociality implications of it. The approach for the presentation of various methods is linked to the features extraction from dataset of experimentations, a large section presents the different modalities and types of data that could be used to find emotions occurrence in individuals (video, images, audio, biological signals etc.). Then AI most used methods (classifier algorithms) for learning scopes will be presented: Machine Learning and its main steps and a particular focus on Deep Learning and Facial Expression Recognition (FER) technologies, trying to find the best situations where one or another shall be used based on dataset types and different modalities of experimentations. The use of AI leads to the topic of correlation between man and machine: since emotional topic is one of the few bridges between them to be fully explored yet and this could represent what is missing by computer to be considered equal to human being, nowadays computer and robots are not capable of proving emotions such animals do and for this is needed learning and cognitive mechanisms peculiar of human beings to be implemented and computable in some way.