Object and face perception Flashcards
What is object and face perception
- This is where we look at objects (everything) and change this to meaningful information
Detection and recognition
- Our brain forms groups of objects based on sounds
- Our brain can recognise that something is there and then uses our memory to decide what is it
- Colour and orientation and spatial frequency also help with feature detection
What is grouping
- Our brain is able to automatically sort features.
For example, proximity, similarity and closure
What happens at a luminance boundary
- Excitation and inhibition cancel each other out when luminance is equal.
- However, at the luminance boundary, there is no balance which means that perceived brightness is increased
Contours for visual objects
- Different neurons respond to different frequencies (high and low) for example, Mona-Lisa
- Vital for turning 2D objects into 3D objects
Shape from shading
- Shading shows texture and depth which helps to identify 3D objects
The role of expectation
- What we expect to see influences what we do see
- ‘built in expectations’
- Everyone has different experiences and expectations
Face detection
- First step is to detect a face this means we have to look at different features of the face that are common across all faces
- Template matching is a simple example of detection
Face recognition
- After detection of a face, features are analysed to categorise
- For example, different facial expressions help to see the emotion
- Also can identify who it is
Challenges of face recognition:
- There are factors which impact what we see; lighting, different context
Are faces special
- Newborns can detect face of their mother
- we have tendencies to see faces in objects
- Pareidolia, we imagine there are faces when there aren’t for example in stones or leaves
Evidence for innate ability
- New born studies which look at preferential looking
- New borns look longer at features which are arranges as a face rather than jumbled features.
Johnsen et al (1991)
Face inversion task (Yin, 1969)
- Better at recognising faces the right way up rather than upside down.
- Much more significant for faces than anything else.
i.e. context is important
Negative Contrast faces (Kemp et al, 1990)
- Harder to recognise when pattern of light and dark changes
- Easier to recognise when pattern of colouring is what we are used to seeing.
i.e. context is important
Hollow Face illusion
- Like as mask you would wear
- Painted inside of the mask (indent side) in a way you would paint a face, makes it seem as though the face is popping out at you
- Painting the face to look as though light is hitting it the way you’d expect although works without highlights.
- Blurred lines between perception and cognition as cognitively you know it’s a mask but you can’t override the 3D affect
- Shapes and shading is a bit ambiguous if you don’t know where the light sauce is coming from
The affect of inversion on the Hollow Face and the Hollow Potato illusions (Hill and Bruce, 1994)
- Made a potato hollow
- Measure how close people had to get to the object until they see it not as convex (popping out) rather than concave (dimple)
- The closer you are, the illusion went away, potato
- You would have to get closer for the face illusion
Do we process faces in a particular way
- The idea that faces are convex overrides knowledge
- Potato illusion shows that we have that same expectation for all objects not just faces
- However, illusion is stronger for faces
Face recognition is different to object recognition Moscovitch et al.1997
- Patient CK finds it hard to recognise objects although they do not find it hard to recognise faces
Prosopagnosia
Meadows, 1974; Damasio et al. 1982
- inability to recognise certain faces
- Although their object recognise is absolutely fine
- Damage to Temporal areas rather than visual areas.
Face specific areas
Perrett et al. (1979, 1982)
- Cells in the superior temporal sulcus are selective for faces. (in monkeys)
- They respond more to human faces than monkey faces
Objected Centred vs Viewer Centred
Hasselmo et al (1989)
- Some neurons respond to faces at different angles
- Some respond to animations rather than photos (respond to eye gaze)
- Object vs Viewer, some neurons only recognise from a particular angle but others can respond from any angle
- Not viewer centred, doesn’t matter where we are to the object.
- Memory is important as well as scocial attention
Functional Imaging (fMRI)
- Fusiform face area responds more to faces than other common objects
- Haxby et al. 2000, should be looking at other things as well such as eye gaze and different angles may activate FFA differently
- People who can’t recognise faces have damage to the fusiform area
Features of Configuration
- There are different features of the face i.e. nose, mouth, eye etc.
- Humans are sensitive to how the face is arranged.
Holistic Processing
Tanaka & Farah 1993
Comparison of faces and housing
- trying to recognise a nose and a front door on it’s own is hard
- if you put the front door on the house and the nose on the face it is more recognisable
- Although, still easier to recognise the face rather than the house
Composite face effect
- Two faces put together (bottom and top half diff)
- Holistic processing is natural
- Can’t recognise the face when bottom and top half are in one face but when split apart, it’s easier to recognise the face
