1. Theories of Perception Flashcards
What is perception?
the distal stimulus (real) –> the proximal stimulus (retinal) –> perceived image. The transduction of signal from sense organs to the brain
Why is the ‘eye as a camera’ idea naive? 3 answers.
- the retina is flat and curved, but we can see straight lines
- the eye constantly makes saccades and updates the image (50x/s)
- the eyes also compensate for body movement (visuo-ocular reflexes)
Define the Gestalt/School Approach
- the whole is greater than the sum of it’s parts. Brain sees the elements and puts them together to form a whole object.
- brain uses heuristics to sort the incoming information (top down)
- also bottom up
Explain the concept of i) similarity, ii) continuation and iii) proximity
i) similar things grouped together (shape, colour, orientation, size)
ii) connected points are seen to belong together, following the smoothest path. Can lead to reification (more spatial info than actually present)
iii) things near to eachother are grouped
Explain the concepts of i) connectedness, ii) closure and iii) common fate
i) physically connected things are a unit
ii) a closed figure is perceptually preferred to an open figure eg square vs cross
iii) things that move in the same direction are grouped
Explain the concepts of i) familiarity, ii) invariance
i) meaningful or familiar looking things are grouped together
ii) objects will be recognised regardless of their orientation, scale or translation (this is something computers cannot do). eg a chair looks like a chair from upside down
Explain figure ground segregation: what is it and what affects it?
- we instinctively perceive things to be in background or foreground
- affected by symmetry, convexity (more convex = foreground), area, orientation, meaning/importance
5 issues with Gestalt approach
- doesn’t account for parallel and unconscious processing in the brain
- some laws explained incorrectly
- some laws do not provide explanation
- Pragnanz: brain will make complex/ambiguous shapes as simple as possible: this is vague
- states the obvious
- using flat 2D images is unfair as we see things in 3D and could change perspective to better view
Define gibson’s ecological approach
- bottom up
- perception is direct, not mediated by cognitive processes
- the information we perceive is enough to interact with environment (goal is the action we are going to do)
- studied in environment, not lab
Active optic flow
- movement of observer allows extra information about objects - what they are and where they are
- in optic flow, the world expands towards and contracts away but can be influenced by active movement
Ambient optic array
light is reflected differentially by textured surfaces and can change with observer movement. There is visual info in reflected light that does not need complex cognitive processing.
What are Gibson’s invariants? Examples of them.
unambiguous information about the environment which can be directly perceived
- examples: the horizon ratio - same proportion of objects are above and below the horizon line
texture gradients - changes in texture can inform us about distance, curves and orientation
How did Gibson and Bridgeman 1987 provide support for Gibson’s ecological theory?
found that participants could use spots on an objects surface to identify the object, state, colour and spatial orientation
What 2 cues does observer movement produce?
1) motion parallax: when the observer is moving, closer objects move faster than far objects. Close objects pass us, far objects seem to follow us.
2) optical flow/expansion: combines motion parallax and retinal size
Define affordance (Gibson’s ecological approach)
- what the object can offer us: can we eat it, grab it etc
- primitive, doesnt require memory or experience
Issue with Gibson’s approach
It is vague: doesn’t acknowledge how we pick up and process info, ignores top-down/neuroscience
Marr’s information processing approach
- bottom up
- computational process of perception
- how we reconstruct elements to perceive whole image
First stage of retinal image analysis in Marr’s approach
The retinal image
Grey level description
produced by photoreceptors. measures the light intensity at each point in the image
Primal sketch
contrast changes (resulting in blobs, edges etc) identified via Gaussian blurring: intensity changes can still be seen at 2+ levels of blurring and are then assigned to edge, bar, termination or blob
2.5D sketch
the orientation, depth and colour relative to the observer
3D representation
object representation independent of the observer
The Constructivist Approach
perception requires past knowledge, experiences and memory.
involves the likelihood principle and unconscious inference
unconscious inference
proposed by Helmholtz. Certain interpretations are more likely to be true, and these are what we will perceive (like brains reflex)
The hollow mask illusion can be explained by this idea (stored knowledge of faces being convex leads to incorrect perception)
Unconscious inference
