Investigating the process of emotion recognition in immersive and non-immersive virtual technological setups

This paper investigates the use of Immersive Virtual Environment (IVE) to evaluate the process of emotion recognition from faces (ERF). ERF has been mostly probed by using still photographs resembling universal expressions. However, this approach does not reflect the vividness of faces. Virtual Reality (VR) makes use of animated agents, trying to overcome this issue by reproducing the inherent dynamic of facial expressions, but outside a natural environment. We suggest that a setup using IVE technology simulating a real scene in combination with virtual agents (VAs) displaying dynamic facial expressions should improve the study of ERF. To support our claim we carried out an experiment in which two groups of subjects had to recognize VAs facial expression of universal and basic emotions in IVE and No-IVE condition. The goal was to evaluate the impact of the immersion in VE for ERF investigation. Results showed that the level of immersion in IVE does not interfere with the recognition task and a high level of accuracy in facial recognition suggests that IVE can be used to investigate the process of ERF.