A - Face Recognition Flashcards
What is face recognition?
understanding and interpretation of faces
it is a form of pattern recognition
Where have theories of face recognition come from?
adult face perception studies
studies of people with impairments in face perception
Why study faces?
faces are central to the human world
it is thought that face perception is different to object perception
central for social interaction, media, imagination etc.
Jeffrey and Rhodes
“faces convey a wealth of information”
What do faces show?
Jeffrey and Rhodes - proportions and expressions are important in identifying origin, emotional tendencies, health and some social information
Little and Perrett, 2007 - can access personality attributions i.e. introvert or extrovert - although this appears easier in women than men
Little et al., 2013 - accuracy in discrimination of self-reported co-operators using static facial information
Specialised mechanisms
these allow different parts of the face to be identified separately and then be processed as whole too
Thompson, 1980
Yin, 1969
Thompson, 1980
the Thatcher illusion shows that each feature of the face is analysed independently and coded relative to gravity - special mechanisms within the brain allow this to happen
looking at different features looks okay but specialised mechanisms quickly realised when upright that it is wrong
Yin, 1969
when faces are upright they are processed by special mechanisms in the right hemisphere
faces presented upside down do not stimulate this mechanism so are treated like objects
highlights different processes for object perception to face perception
Brain asymmetry
left hemisphere processes info from and controls right side of body
right hemisphere processes info from and controls left side of body
Face expression
more prevalent on the left side of the face (80%)
explanation = emotional/motor control bias to right hemisphere
Prosopagnosia
helpful in understanding how normal face perception may work
Bodamer, 1947 - brain damage can produce problems in face recognition, even one’s own reflection
Farah, 1990 - 94% of prosopagnosia patients have experienced damage to the right hemisphere of the brain
Recognition
humans are really good at recognising faces even with different angles, lighting etc.
but humans are really poor at recognising unfamiliar faces
computer analysis systems are even poorer at face recognition - Hancock, Bruce and Burton, 2000
EWT is not reliable because of this poor recognition of unfamiliar faces
Internal and external features
you describe people you don’t know well in terms of external features - i.e. glasses, hair colour
you describe people you do know well in terms of internal features - i.e. personality
Information Processing Models - Bruce and Young - description
1986 - model of face processing
series of stages each accessed serially
face is seen, the structural encoding begins
view-centred descriptions - expression analysis, facial speech analysis and directed visual processing - all refine information given from structural encoding
expression-independent descriptions assess FRUs, PINs and name generation
all assessed in cognitive system where information is stored
Bruce and Young, 1986 further description
the two paths
1 - recognition of faces - has a feature-by-feature analysis - this creates a structural model of a face which can be compared to faces stored in memory - explains why faces of known people can still be recognised at novel angles, lighting etc.
2 - recognition of expression - looks at simple physical aspects of the face i.e. age, gender, facial speech
Bruce and Young, 1986 - evaluation
the cognitive system in the model is very vague and its processes are not explained
Thompson, 1980 - supports notion that facial features are processed separately but challenges the idea that this works for non-upright faces (can also bring in Yin, 1969)
Bruce and Young, 1993 - support the model - looked at soldiers with brain damage - found face processing was modular
Case study of Mr W - couldn’t recognise faces but could do other things i.e. recognise a nose - highlights modularity in face perception
Information Processing Models - Burton, Bruce and Johnston, 1990 - description
Interactive Activation and Competition (IAC) Model
lots of priming studies resulted in a tweaked model
people faster at recognising certain faces after primes
FRUs activated by visual information which activates PINs to produce recognition as a lexical concept (name)
Additional pools of semantic information - SIUs, which work synergistically with NRUs and WRUs to provide semantic information; being a connectionist model it simulates the excitatory activation and inhibitory links between the pools.
Burton, Bruce and Johnston, 1990 - evaluation
Supports sequential nature of facial recognition
Supports repetition/semantic and cross-modal priming
Supports neuropsychological data; prosopagnosia and Capgras delusion- can’t recognise the face but can recognition expressions - separate processes
What does structural encoding consist of?
featural processing - i.e. looking at distinct features of that person
- Bradshaw and Wallace, 1971 - pps identify differences in mug-shots much quicker when more differences are present between features - suggests sequential processing
configural and holistic processing
- Tanaka and Farah, 1993 - pps learnt faces then tested on individual features in normal and scrambled faces - features better recognised when presented within a whole face
Robust processing
we can still recognise faces well when they have undergone robust processing
i.e. vertical stretching, shearing, inversion
Multidimensional face space
Valentine, 1991
easier to recognise those with distinctive traits because there are fewer people in the memory to compare them against
closer towards the average = more difficult to recognise them because there are so many others who could also fit
- this can also be linked to Bruce and Young’s model of face processing - specifically the face recognition unit - how quickly faces are recognised may depend on this
Adaptation
links to aspects of faces and people perceiving them that influence face preference
facial adaptation
Meissner and Brigham, 2001 - better recognising own ethnicity better than other ethnicity groups
better also at recognising those in your in-groups than your out-groups
there is a large literature on exposure biasing subsequent perception in faces
Categorical perception
links to aspects of faces and people perceiving them that influence face preference
how similar things look different depending on whether they are classified as the same category or not, i.e. male or female but still both human
Fugate, 2013 - many studies show that adult humans show categorical perception for human emotional faces
- perceptual and conceptual knowledge interact to explain how people see discrete emotions in others’ faces
Category contingency
links to aspects of faces and people perceiving them that influence face preference
experience of male faces has limited influence on perception of female faces
experience of female faces has limited influence on perception of male faces
implies the brain holds distinct representations of male and female faces
Prototype referenced coding
links to aspects of faces and people perceiving them that influence face preference
we can imagine a multidimensional ‘face space’ (Valentine, 1991) where each individual face is a point in space with a theoretical average of all faces at the centre
this average face my be useful in recognition of identity
Appearance
links to aspects of faces and people perceiving them that influence face preference
- readily accessible and impacting on first impressions of people
- as we get to know people, appearance plays less of a role in what we think about them - we judge on appearance because at the time it is a shortcut for judgement
- Willis and Todorv, 2006 - snapshot picture of someone can quickly lead to formation of judgement for trustworthiness, attractiveness, competence, likeability, aggressiveness
- neotenous features not just present in human children but also in some animals - we respond to them and their faces - perhaps we are hard-wired to respond to these types of features - automatic responses to infantile cues
Infant face processing
links to aspects of faces and people perceiving them that influence face preference
infants have a preference for human faces over objects
it is also suggested that infants have a specialism at upright faces
also thought that women are preferable over men for babies - perhaps due to connection with mother
Slater et al., 2000 - show similar preferences to adults in more preferable faces to look at
Attractiveness
links to aspects of faces and people perceiving them that influence face preference
similarity to face population mean influences its attractiveness
average appearance may indicate diverse mixture of genes - genetic diversity leads to stronger immune system
Evolution and Attractiveness
can these preferences also be linked to evolutionary theories of survival?
some factors will be linked to benefits to the choose no matter the environment or person
this would lead to selection and so universal preferences for some traits
e.g. no culture would prefer a sick or diseased face when choosing their partner so it is expected that in all cultures health people will be preferred when choosing a mate
Masculine and feminine male faces
links to aspects of faces and people perceiving them that influence face preference
Perrett et al., 1998
masculine males seen as dominant, low warmth and bad parents
feminine males seen as honest, warm and good parents
masculinity may be preferred for protection of offspring but femininity is important for investment of resources in offspring
Little et al., 2010 - masculinity in faces should be preferred in a woman’s life when reproduction is most relevant
Repetition priming
Semantic priming
Definitions
Repetition priming - face becomes familiar on repeated viewing
Semantic priming - recognised face carries implicit semantic information for recognition of another face
FRUs
PINs
Name generation
face recognition units (FRUs - looking at stored memory for faces using structural features), person identity nodes (PINs - specific info about that person) name generation (this may take a bit more time)
Attractiveness in face perception studies
Langlois et al., 1991 - composite images, the more faces that go into an image the more attractive it becomes - it is closer to the average
symmetry is also found attractive - higher preference for it
Penton-Voak et al., 2001 - ratings of women correlate with measured symmetry
Perrett et al., 1999 - symmetry is found attractive in both males and females
