11: Facial Recognition

Facial recognition involves the ability to remember and identify faces.
Key inquiries surrounding facial recognition include:
- What constitutes facial recognition?
- How do we process and evaluate it?
- In what contexts is it applicable?
- What factors influence facial recognition capabilities?

Facial recognition is an innate ability rooted in evolution (factor of pareidolia, survival instinct).
Different facial features contribute unequally to facial recognition:
- Internal Features: Includes eyes, nose, mouth, which are prioritized in recognition.
- External Features: Includes ears, cheeks, chin, and hair, which are less emphasized.
Age influences how facial features are recognized and processed.
Recognition can involve both local features (specific parts) and configural aspects (arrangement of features).

There are two primary processing strategies:
- Holistic Processing: Viewing faces as a complete unit.
- Feature Processing: Recognizing faces by analyzing individual components.
The encoding of faces typically occurs as a single entity, contrary to object recognition which focuses on parts.
This concept is informed by the following effects:
- Inversion Effect: Difficulty in recognizing inverted faces compared to upright faces.
- Composite Effect: Challenges encountered when trying to recognize faces that are combined with other faces.

Proposed by Valentine (1991), posits that:
- Faces are represented on a multidimensional axis where different axes correspond to facial features.
- The position of a face in this space is determined by a composite of local and configural features.
- The concept emphasizes the ‘averageness’ of faces, where central, average faces are more easily categorized and recognized.

Suggests that metric distances among facial features are crucial but vary significantly across individuals.
Conducted tests on memory for faces with small changes to their configural features and recognized limitations in metric stability based on varied encounters with the same face.

Participants: 24 subjects.
Stimuli: 120 celebrity faces (both familiar and unfamiliar).
Manipulations: Adjustments to distances between facial features (large and short distances between eyes).
Presented original and altered pairs, including inverted pairs, across 240 trials (Two-Alternative Forced Choice - 2AFC).
A familiarity check was performed at the end.

Participants: 36 subjects.
Tasks:
- Familiarity categorization.
- Discrimination task assessing configuration changes.
Subjects viewed one face at a time with a requirement to identify changes in configuration across two blocks of 60 trials.

The studies indicated a sensitivity to changes in the configuration of the eyes.
Noted that such sensitivities have limitations due to the variability of face configurations in different contexts.
Suggests that cognitive representations of faces must be adaptable to accommodate these variations.

Different methodologies employed in evaluating facial recognition include:
- Forced choice paradigms.
- Old/New recognition tasks.
- Ratings tasks assessing recognition confidence.
- Eye-tracking to understand visual attention and focus.