Facial Expression Recognition by 2D Image and 3D Mixed-Reality Image

Facial expression recognition is a fundamental aspect of human communication, influencing social interactions across various environments. With the advancement of Mixed Reality (MR) technology, which merges real and virtual elements, it is crucial to investigate how facial expressions are perceived in MR environments compared to traditional two-dimensional (2D) images. This study aims to examine differences in facial expression recognition between 2D images and 3D MR images, focusing on three key facial expressions: neutral, fabricated smile, and Duchenne smile. Three-dimensional facial images were acquired using a Revopoint MIRACO 3D scanner, while 2D images were captured with an iPhone 12. Participants evaluated the images under two conditions: 2D image presentation on a computer screen and 3D image presentation using Apple Vision Pro with Polycam. Subjective evaluations were conducted using a 7-point Likert scale, assessing perceived naturalness, emotional intensity, and affective impressions. The findings provide insights into how facial perception differs in MR compared to traditional 2D imagery, contributing to research on human-computer interaction and affective computing.