Detection of Social Signals During Communication in Media Training Using Commercial Automated Emotion Recognition Systems

It is well recognised that non-verbal signals play an important role in communication. Observation of videos to understand non-verbal behaviour is time-consuming and limits the potential to incorporate detailed and accurate feedback of this behaviour in practical applications. The aim of the current research is to investigate which combinations of social signals are most promising for evaluating trainees’ performance in a media interview using automatic recognition technology. Non-verbal signals were captured during practice media interviews included facial expression, hand gestures, vocal analysis and ‘honest’ signals. Interviews were categorised into effective and poor communication exemplars based on communication skills ratings provided by trainers and neutral observers. Features were selected using a correlation-based feature selection method and their accuracy was assessed using an instance-based classifier. Results revealed that participants who were rated as effective communicators in the context of a media interview smiled more, showed more vocal signals associated with passion, had a relaxed posture and had constant overall movements. These findings suggest that automated technology is capable of detecting social signals that are important for effective communication in a media interview. Results have implications for the automatic evaluation of media interviews with a number of potential application areas.