Ensemble crowd perception: a viewpoint-invariant mechanism to represent average crowd identity

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computer Vision

Background:

• Ensemble perception enables rapid averaging of visual information for low-level features and high-level objects.
• The representation basis of ensemble perception, particularly viewpoint invariance for complex objects, remains largely unexplored.

Purpose of the Study:

• To investigate whether ensemble perception of faces relies on viewpoint-invariant object representations.
• To determine if the brain integrates facial information across multiple orientations for ensemble coding.

Main Methods:

• Participants viewed sequential crowds of faces with varying identities and viewpoints.
• Ensemble perception of mean crowd identity was assessed using a test face.
• Control experiments manipulated viewing time and sampling strategies.

Main Results:

• Accurate perception of mean crowd identity was achieved despite integrating information from multiple face viewpoints.
• Ensemble coding precision was not solely dependent on viewing duration.
• Observers integrated information from numerous faces, rather than sampling a subset.

Conclusions:

• Ensemble perception operates effectively for high-level object recognition, specifically for 3D viewpoint-invariant faces.
• This demonstrates a high level of abstraction in ensemble coding, extending beyond simple features.
• The findings support the role of viewpoint-invariant representations in ensemble perception of complex objects.