L12 - Face Recognition

Question 1

What is different between facial and object recognition?

Answer 1

• Face recognition to identify individual, not categorise face as a face.
• Faces are recognised holistically
• Have more expertise with faces.

Question 2

What is Prosopagnosia?

Answer 2

Deficits in face recognition, with intact object recognition.

Question 3

What are the configurable properties of faces?

Answer 3

• First-order relations between features: eyes above nose, nose above mouth
• Holistic processing: all features perceived together.
• Second order relations between features: exact metric differences between features. 1

Question 4

What is evidence of the holistic processing of faces?

Answer 4

• Part-whole effect
• Face inversion effect
• Prosopagnosia: better at discriminating features in isolation.
• Composite effect

Question 5

What is the Part-Whole Effect?

Answer 5

It’s easier to recognise a face feature when it’s presented within a face.
E.g. when learning Larry’s face, it’s easier to determine which one is Larry, not just his nose individually. If trained on a scrambled face, it’s easier to make judgements on isolated feature.

Question 6

What is the Face Inversion Effect?

Answer 6

• Upside-down faces are harder to recognise (slower and less accurate)
• Parts are hard to discriminate in inverted faces.
• Whole faces are easier to identify when upright.

Question 7

What is the Composite Effect?

Answer 7

It’s difficult to judge whether the top half and bottom half of a face are from separate people, because we perceive it as one face.

Question 8

What is the Other Race Effect?

Answer 8

• Some evidence that faces of other races are processed less holistically.
• Reduced inversion effect, composite effect findings mixed.
• Some evidence that holistic and part-based information more effectively processed in own race.

Question 9

Why makes familiar face recognition different?

Answer 9

Burton argued that configural information can’t explain ability to recognise familiar faces because different lighting, postures etc., you can still recognise.

1. Large configural changes leave recognition unharmed.
2. Non-configural changes harm recognition. E.g. can’t recognise people in photo negatives.
3. Face Shape vs. Face Texture. Put familiar face onto another structure, and can still identify. Texture dominates.

Question 10

How do we recognise familiar faces (Jenkins & Burton, 2011)?

Answer 10

Abstract statistical averages of all the instances of the familiar face, and is stable from contamination.