Perception + action wk 1 Flashcards
4 approaches of how we gain visual perception from retinal info
Gestalt approach
Gibson’s ecological theory
Marr’s info processing theory of perception
Constructivist approach to perception
what do the 4 approaches differ in
Diff approaches to
- Bottom up vs Top down processing
- Goal of perception
- Methods of study
Two approaches to perception of form and organisation
Marr’s approach, concerned with contrast change
Gestalt approach, concerned with rules of perceptual organisation
Marr’s approach to perception of form and organisation (process)
Retinal image is analysed sequentially at different levels:
Retinal image –> Grey level description ( measure intensity of light at each point)
—> Primal sketch (representation of contrast change e.g. blobs, edges, bars etc)
–> 2 1/2 D sketch ( representation of orientation, depth, colour ,relative to observer,) Its not 3D because it is observer-orientated. (unseen parts of scene
and objects)
—> 3D representation- representations of objects independent to observer
Gestalt laws of perceptual
organisation
- similarity
- Good continuation (lines follow smoothest path)
- Proximity (close = grouped)
- Connectedness (physically connected = unit)
- Closure (e.g 4 on a dice looks like a square)
- Common fate (moving together in same direction = grouped)
- Familiarity ( once discovered one indicator of object, don’t need all)
- Invariance (recognise under diff visual situations)
- Pragnanz- “good figure” (central law- most simple + stable shape)
Figure-ground segregation- properties effecting whether area is seen as figure or ground are:
(background or foreground)
- symetry (symetry usually figure)
-Convexity (convex usually figure)
-Area (smaller area usually figure)
-Orientation (vertical + horizontal usually figure
-Meaning/importance. this implies top-down processiing occurs here
Gibson’s ecological theory of perception
- Bottom up approach
- no diff between sensation and perception
- sufficient info in retinal array. retinal image provides rich info sufficient for interaction
- complex cog processes unnecessary
- Ecological: study in natural environment, not the lab (no 2D displays)
- importance of motion, we move around. movement gives us rich info on what’s going
- Argued observer is ACTIVE, movement of observer provides additional stream of info.
- we have an Ambient optic ray
invariants
example 1 horizon ratio
- Unambiguous info ab envi
- can be directly percieved
- e.g. horizon ratio relation = proportion of object above horizon is constant with changes in distance but not size
Invariants example 2
Texture gradients
- Unambiguous info ab envi
- can be diretly percieved
- changes in texture in the optic arrray tells us about distance, orientation and curvature of surfaces
Optical flow is a combo of
combo of motion parallax and retinal size
Gibson’s idea of affordances
can recognse purpose of object without recognising the object itself
enough info in visual image to automatically know what to do with it.
= intrinsic affordances
Constructivist approach to perception
Seen as a compromise position
2 things can agree on
- retinal image does not provide sufficient information
- perception depends upon stored knowledge
Argued Fixed unconscious neural processing.
- Illusions: impervious to experience
- Naive optics (e.g. movement of sun in sky)
- perception depends upon stored knowledge
- Retinal image dies not provide sufficient information
Generation of the perceptual hypothesis
Richard Gregory
constructivist approach
many illusions explained by stored knowledge leading to inaccurate perceptual hypothesis
(so we kind of use our knowledge to determine what we are most likely to be looking at)
