Familiar Face Recogniton: Cognitive Psychology Flashcards
Davies, Ellis, and Shepherd (1978)
10 celebrity face photos; name them. 93% accuracy.
Within-Person Variability
Appearances can change in lots of different ways, but we can still recognize people.
Within-Person Variability study
Jenkins, White, Van Montfort & Burton (2011). Used 20-40 photos with one person. 100% accuracy.
Recognition From Poor Quality CCTV (1)
Burton, Wilson, Cowan & Bruce (1999). Glasgow University. 95% familiar face recognition accuracy.
Recognition From Poor Quality CCTV (2)
Bruce, Henderson, Newman, & Burton (2001). Is it the same person in the photos as in the CCTV image? 97% familiar face recognition accuracy.
What are criticisms of Bruce, Henderson, Newman, and Burton (2001)
Had a between-subjects design (Glasgow University students familiar with Glasgow University lecturers, Paisley Uni students not familiar). Difficult to measure familiarity (to what extent did each person have familiarity towards lecturer).
Recognition From Corneal Reflections
Jenkins & Kerr (2013). Reflection of the bystander in an eye photo. 84% familiar face recognition accuracy from corneal reflections.
What is a criticism of Jenkins and Kerr (2013)
The corneal reflection photos require high quality photos (face facing forward, focused, good lighting)
Recognition From Pixelated Images
Lander, Bruce, & Hill (2001). 5-second presentation of celebrity faces, to name them. Pixel: Familiar face recognition highly accurate, but motion reduces. Blur: moving images recognized more accurately than static.
Artificial Distortion: Altered Images
Hole, George, Eaves & Rasek (2002). Name the celebrity task, RT. Artificial distortions had no effect on the recognition of familiar faces.
Argument Against Configural Processing
Maurer, Le Grand & Mondloch (2002). Configural processing is the ability to perceive the spatial relationships between the features of a face. Destruction of these relations by distortion, for example, leaves recognition unharmed.
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 1
You encounter a person you know. Use the visual system to see the face.
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 2
Face representation in the brain. We create a neural representation in our brain. We recognize it is a face we are seeing.
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 3
Structurally encoding the face. Send information to certain parts of the brain (structural codes). Abstract identity-based structural code (gender, expression).
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 4
Activates face recognition units. No names yet or semantic information. Signals to other processes to find out who it is. Each person we are familiar with has an FRU and they have their own structural codes. FRU’s are stored in memory with their own structural codes. “This person is familiar… but who are they?”
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 5
Activate person identity nodes (PIN). Links recognized face to semantic information about the person.
The Bruce & Young (1986) Model of Familiar Face Recognition: Step 6
Recall the person’s name. It is a complementary process, one step after another, in order to find the name and recognize the face.
Evidence to support B&Y model (1)
Young et al. (1986b). Familiarity decisions (FRU stage) are made faster than semantic decisions (PIN stage), which supports the model. In Task 1, participants quickly identified if faces were familiar or unfamiliar, testing recognition at the FRU stage. In Task 2, they identified if faces belonged to politicians or actors, testing semantic recognition at the PIN stage. Familiarity recognition was significantly quicker than extracting semantic information, supporting the model’s claim that the familiarity stage (FRU) occurs before the semantic stage (PIN).
Evidence to support B&Y model (2)
Young et al. (1986c). Participants took longer to make a politician/popstar semantic categorization when presented with the person’s name, in comparison to when they were presented with a face. The model says that face recognition (via FRUs) is faster, while name recognition relies on later stages (PINs and semantic processing), explaining the delay.
What is a piece of evidence against the Bruce and Young model (1986)
Calder and Young (2005) say expression and identity are more integrated than the model suggests. Ganel et al (2005) found neural imaging evidence showing the FFA processes both simultaneously
What could the FRU held in memory be?
An ‘Average’ Face for each person we know.
FRU – A Face Average? Study 1
Burton et al. (2005). Lots of photos of Harrison Ford throughout his career merged together photos to create an average face. Name shown then face average or normal photo shown. Then yes/no if match. Faster responses when averages were used. Evidence that our FRU representation could be face averages (faster response time = closer match to the FRU representation).
FRU – A Face Average? Study 2
Koca and Oriet (2023) found that averaging faces’ within-person variability over time leads to face familiarisation. People become familiar with faces by averaging different views of the face.
FRU’s: A Face Average? Machine studies (1)
Jenkins & Burton (2008). Website that when you upload a photo, it shows your celebrity look-alike. When uploading a photo of a celebrity (standard image), it matched to the same celebrity 54% of the time. Then when uploading a face average, it matched the same celebrity 100% of the time.
