theories of visual perception Flashcards
what are gestalt’s laws of perceptual organisation
- similarity
- good continuation
- proximity
- connectedness
- closure
- common fate
- familiarity
- invariance
- Pragnanz - good figure
similarity
similar things appear to be grouped together
grouping can occur due to shape, lightness, hue, orientation, size…
good continuation
points that, when connected, results in straight or smoothly curving lines, are seen as belonging together
and lines tend to be seen in such a way as to follow the smoothest path
reification
the experienced percept contains more explicit spatial information than the sensory stimulus on which it is based
proximity
things that are near to one another appear to be grouped together
connectedness
things that are physically connected are perceived as a unit
closure
of several geometrically possible perceptual organisations, a closed figure will be preferred to an open figure
we tend to complete a broken figure because of the strong closure cue for organising what we see
for which rules does reification occur
closure
good continuation
common fate
things that are moving in the same direction are grouped together
objects with the same orientation are grouped together
familiarity
things are more likely to form groups if the groups appear familiar or meaningful
invariance
simple geometrical objects are recognized independent of rotation, translation, and scale; as well as several other variations
which rule of gestalt is a serious problem in computer vision
invariance
figure-ground segregation basics
normally in a visual scene some objects (figures) seem prominent, and other objects recede into the background
what does figure-ground segregation infer
top down processing
properties that affect whether an area is seen as figure or ground
symmetry
convexity
area
orientation
meaning/ importance
symmetry figure ground segregation
symmetrical areas usually figure
convexity figure ground segregation
convex shapes are usually figure
area figure ground segregation
stimuli with comparatively smaller area are usually figure
orientation figure ground segregation
vertical and horizontal orientations usually figure
meaning/ importance figure ground segregation
meaningful objects more likely to be seen as figure
implies attention: top-down process
problems with the gestalt approach
underplay the parallel processing and unconscious processing the brain does
explanation of how some of their laws worked was wrong
their laws provide a description of how things work rather than an explanation
some of the laws are ill defined
-e.g. pragnanz: what is the simplest and most stable shape?
stating the obvious?
pragnanz
when people are presented with complex shapes or a set of ambiguous elements, their brains choose to interpret them in the easiest manner possible
distal stimulus
object in the real world
proximal stimulus
representation of the stimulus on the retina
gestalt psychology
the whole is greater than the sum of its parts - max wertheimer
segregation
how we separate figure from ground
grouping
how we group parts of a stimulus together
structuralism
perception is a combination of individual sensations that can be reduced to simple, individual elements
optic flow
the way the world flows and expands toward us, and contracts and flows behind us
what did Gibson notice about optic flow?
that it was active - movement of the observer provided an additional stream of information
ambient optic array
the structured arrangement of light reflected by textured surfaces with respect to a point of observation
-changes due to movement
invariants
unambiguous information about the environment
can be directly percieved
horizon ratio relation
proportion of object above the horizon is constant with changes in distance (of the same size standing on the same ground), but not with changes in size
texture gradient
changes in texture in the optic array tell us about distance, orientation and curvature of surfaces
texture becomes smaller as distance increases
empirical support for Gibson’s theory
Gibson and Bridgeman (1987)
participants could correctly identify objects, state their colour and identify the lighting conditions and spatial orientations just from black and white photos of objects surfaces
the average subject identified ~2/3 of the photographs correctly
motion parallax
things far away move more slowly than things nearby
speed of movement tells us about the distance to the object
monocular cue to depth
motion parallax in animals
used much more by animals that dont have much binocular overlap
head bobbing and orthogonal running
expansion (optic flow) is a combination of
motion parallax and retinal size
practical application of Gibson’s optic flow ideas
horizontal lines on the road painted closer together as the driver approaches junctions. Creates the illusion of increasing speed which causes the driver to slow down
what is an affordance?
an action possibility formed by the relationship between an agent and its environment
memory and experience not necessary?
criticisms of Gibson’s theory of perception
vague? how is this information picked up?
ignores top down processing (experience and memory)
ignores neuroscience
grey level description
measuring intensity of light at each point in an image (each point on the retina)
produced by activation of retinal photoreceptors
primal sketch
representation of contrast change (blobs, edges, bars etc.) over a range of spatial frequencies
2 1/2 D sketch
representation of orientation, depth and colour relative to the observer
3D representation
representation of of objects independent of the observer
raw primal sketch goal
identify the objects edges
primal sketch to 2 1/2 D sketch
primal sketch combined with depth cues, colour, motion
it is not 3D because it is observer oriented (unseen parts of scene and objects)
2 1/2 D sketch to 3D sketch
2 1/2 D sketch analysed for 3D volume primitives (cylinders, cones, cubes etc.)
produces 3D representation that is independent of observer
conscious experience of vision
importance of the computational approach
an algorithm is more likely to be understood by understanding the problem that has to be solved, rather than examining the mechanism in which it is embodies
to understand perception by studying neurons is like trying to understand a bird in flight by only studying feathers
criticism of Marr’s information processing approach to perception
retinal image is not always sufficient to allow reconstruction
role of memory and experience?
the constructivist approach to perception
retinal information does not provide sufficient information
perception depends upon knowledge (memory) and experience
generation of perceptual hypotheses
unconscious inference
involuntary, pre-rational and reflex-like mechanism which is part of the formation of visual impressions
fixed unconscious neural processing
examples of fixed unconscious neural processing
illusions: impervious to experience
naive optics
titchner illusion
brain over emphasises size differences in grouping and separating objects
helmholtz
unconscious inference
likelihood principle
what explains many illusions
explained by stored knowledge leading to inaccurate perceptual hypotheses
what is the hollow mask illusion explained by?
our stores knowledge that faces are convex
positives and negatives of the constructivist approach to perception
pro: highlighted the importance of both bottom-up and top down processing
con: criticised for being vague
