hide
Free keywords:
3D perception; face space; similarity judgements; RSA; encoding models
Abstract:
Faces are socially relevant stimuli that can be distinguished by the spatial arrangements of their visual features. However, face perception has been mostly investigated with static 2D images, which differs from everyday life experience. In an online study, we investigate face perception in two viewing conditions (2D & 3D). We compare the cognitive face space for these conditions, by modeling the acquired human similarity ratings with similarity matrices computed from physical face attributes and feature maps of deep learning-based face recognition models. Lastly, we fit these models to the human similarity judgements to explore relevant facial features between the viewing conditions. Unveiling differences between 2D and 3D perception of faces will further our understanding on the role of stimulus presentation on face processing.
