Perception
§ Perception is the process whereby sensory experience is
organized and interpreted to allow us to experience objects and events as not
just isolated sensations but as things that possess certain invariances such as shape, size,
etc.
§ Central issue of perception: How can we perceive the constant properties of the distal
stimulus, given the inevitable variations
in the proximal stimulus?
§ The two contrasting emphases, nativism
and empiricism, are central to issues in perception
§ Best evidence suggests that some perceptual organization
derives from built-in or innate mechanisms, while some is acquired through
experience
§ Some perceptual organization appears to occur in a bottom-up process while other
organization seems to arise from a top-down
process
Bottom-Up Vs Top-Down Perception
§ Bottom-up perception
§ Perception that consists
of recognizing and processing information about the individual components of
the stimuli and organizing them to form a whole
§ Feature analysis
Top-Down perception
§ Perceive the whole and
then the individual parts
§ Perception that is guided
by higher-level knowledge, experience, expectations, and motivations
Feature Analysis
Example of Top-down Processing
Figure-Ground Distinction & Gestalt Principles of
Organization
§ One of the most basic
organizing processes in perceiving objects and the world around us is the
separation of objects or figures from their background or ground--the Figure-Ground distinction
§ What are the
organizational principles that are used in accomplishing this?
§ The Gestalt psychologists
(Koffka, Köhler, Wertheimer) were very interested in this problem and suggested
a number of principles
§ Note that use of these
principles to blur the figure-ground distinction is the basis of camouflage
Ambiguous Figure & Ground
Yet another example…
Gestalt Organizing Principles
Proximity
Similarity
Closure
Continuation & Symmetry
Depth Perception
§ A major enigma about
sight is the very idea of a 3-dimensional perceptual world
§ The retina is a
2-dimensional screen. How can we learn about (i.e., see) 3 dimensions when all
we have as input are 2-dimensional arrays?
§ To perceive depth, the
visual system makes use of a number of cues
Depth Cues
§ Monocular
§ Linear Perspective
§ Clearness
§ Occlusion or
Interposition
§ Depth of Focus
§ Shadows
§ Motion Parallax
§ Binocular
§ Retinal Disparity
§ Eye Convergence
Linear Perspective (Monocular)
§ Parallel lines
converge
§ Distant objects appear
smaller
§ Textured elements
become smaller with distance
Linear Perspective
Linear Perspective
Clearness (Monocular)
§ Objects farther away
appear less distinct, less saturated, fuzzier and more blurred than those
nearby
§ The more atmospheric
particles between the viewer and a distant object the more light is scattered
Occlusion or Interposition (Monocular)
Shadows (Monocular)
§ When light strikes an irregular surface, certain parts are illuminated, others
cast into shadow
§ Appearance of shadows
tells a lot about depth of parts concerned
§ Note that the brain
makes assumptions about the most likely source of light to help interpret the
information
§ Cast shadows aid in
locating an object
Accommodation (Monocular)
§ The change of shape
performed by the eye lens to focus on an object aids the brain in determining
the object's distance.
Motion Parallax (Monocular)
§ When we move or the
environment moves, nearby objects' images move more quickly across the retina,
far away objects move more slowly
§ One of the most
important cues
Retinal Disparity (Binocular)
§ Because the two eyes
are spaced apart, when both look at the same object the images obtained are
slightly different. The nervous system fuses the information to perceive depth
§ Sometimes called stereoscopic vision
§ Example: "Magic
Eye" pictures: 2-D pictures become 3-D when viewed with eyes converged
behind the poster
Convergence (Binocular)
§ The angle between the
line of sight of each eye is larger as an object moves closer (cross-eyed
effect)
§ Note this is different
from, but works together with, accommodation (a monocular cue) especially for
nearby objects
Movement Perception
§ What is it that causes
us to perceive that something has moved?
§ The phenomenon that
probably throws the most light on how we perceive motion is the Phi Phenomenon (or Apparent Movement or Stroboscopic Movement)
§ There is a vast
literature in psychology on this phenomenon
Phi Phenomenon
§ Simplest case: Two
lights are switched on and off so that just after one light has gone off the
other comes on. What is seen--provided the distance
between them and the timing is right--is a single light moving from the 1st
position to the 2nd
§ The simple
intermittent displacement of a stimulus (if distance and timing are correct) is
sufficient to signal movement
§ Interesting web site
on Phenomenology of Motion: http://www.cis.rit.edu/people/faculty/montag/vandplite/pages/chap_13/ch13p2.html
§ Phi phenomenon and persistence of vision are the basis of
perception of movement in TV and movies
Persistence of Vision
§ Persistence of vision
is the inability of the retina to follow and signal rapidly changing
intensities
§ A light flashing at greater than about 50 flashes/second appears
steady--though for bright lights and peripheral vision the critical fusion frequency may reach 100
flashes/second
§ Computer monitors and
TV screens refresh the screen many times/second, but if the refresh rate is
high enough the intensity appears steady; if the refresh rate is too low,
annoying flicker becomes apparent
§ Flicker can be annoying
and even dangerous to sufferers from some forms of epilepsy (photosensitive
epilepsy)
§ A highly flickering Pokemon cartoon in
§ The exact mechanisms
underlying the negative effects of flicker are not well understood
Other Movement Effects
§ Waterfall aftereffect: http://www.yorku.ca/eye/mae.htm
§ Spiral aftereffect:
http://dogfeathers.com/java/spirals.html
§ Fechner color illusion: http://dogfeathers.com/java/fechner.html
§ There is no definitive
explanation for this phenomenon. One possible explanation is based on the
finding that short wavelength or blue cones respond more slowly to light than
red or green cones
§ Induced movement: A spot of light is
projected on to a screen which is moved. It is the stationary spot which is
seen as moving
§ Explanation: It's generally
smaller objects that are most likely to move, so the brain appears to take the
best bet when the situation is ambiguous, and assumes the small object moved
Are Perceptual Processes Built In or Data Driven?
§ There are a number of
perceptual processes that appear to be built in, and a number that appear to
strongly depend on experience
§ Several lines of
evidence used to examine this question
§ Evidence from studying
very young (newborn, newly hatched) organisms
§ Studies of recovery of
function after sensory deprivation
§ Neurophysiological
studies of function of single cells in the brain
Studies of Very Young Organisms
§ Do we have to learn
how to see or are we born with the ability? William James described the world
of the infant as a 'blooming, buzzing confusion," but is this so?
§ Ability to
differentiate between different visual stimuli is inferred from consistent
differences in response to them.
§ Fantz's
studies of newly hatched chicks showed that they have a built in preference for
pecking at round shapes (similar to grain or seeds) rather than other forms
(such as small pyramids)
§ Babies seem to have a
built in tendency to focus attention on edges--especially corners--of visual
forms
§ Both babies under 2 days old and 2-5 days old looked roughly 3 times as
long at a black and white picture of a face as at a plain colored circle. This
early preference for facelike forms argues for a
built in mechanism
§ Fantz's
experiments showed infants show preference for facelike
forms even when all the features of a face in jumbled form were presented as an
alternative
Infants' Perception of Faces
§ Fantz
(1961, 1963) found that infants preferred patterned figures, such as faces over
plain stimuli
§ Infants made the
distinction between a correct schematic face and when the elements of the face
were scrambled
Visual Cliff Apparatus
§ Used to study
development of depth perception in humans and animals
Results of Visual Cliff & Related Studies
§ At ages tested (6-14
months) human infants avoided deep side. Many cried when mother called or
crawled away from her--unwilling to cross the chasm. Some patted the glass but
still backed away
§ Experiment doesn't
prove that infants' avoidance of chasm is innate, but evidence from similar
experiments with newborn animals does suggest that both depth perception and
fear of height is either inborn or develops very early. Even animals that are
able to stand and be mobile within hours of birth avoid the chasm as soon as
they are able to stand and walk
§ On the other hand,
animals (cats & rats), who are reared in the dark, fail to gain the ability
to discriminate depth differences. However, as they gain experience in a
lighted world, their depth discrimination ability usually improves rapidly
& eventually reaches normal levels
§ Experiments with 3-D
perception suggest that human infants begin showing stereoscopic depth
perception at about 4 months--appears to be due mainly to brain maturation
rather than learning
§ Bottom line: Experience & innate
factors interact to produce an animal's ability to perceive depth
Perceptual Ability After Sensory
Deprivation
§ The question of how a
man born blind who had his sight suddenly restored as an adult would perceive
the world is one that has fascinated people for centuries. The problem was
considered by Locke and Berkeley who argued that we should not expect such a
man to know that anything was:
§ "high or low,
erect or inverted . . . for the objects to which he had hitherto used to apply
the terms 'up' and 'down', 'high' and 'low,' were such only as affected or were
some way perceived by his touch;
but the proper objects of vision
make new sets of ideas, perfectly distinct and different from the former and
which can by no sort make themselves perceived by touch." (
§ Now there have been
quite a few cases of recovery of sight after being born blind, but only a few
have been adequately described or studied. Some of the cases are as the
empiricists expected, i.e., the person sees little at first, unable to name or
distinguish between even simple objects or shapes
§ On the other hand,
some individuals did see quite well almost immediately, particularly those who
were active and intelligent
Restored Vision as an Adult
§ Examples
§ R.L. Gregory's patient
known as S.B.
§ Shirl
Jennings, whose case inspired the movie, 'At First Sight'
§ Perception of distance
often seems peculiar. S.B. thought he would be able to touch the ground below
his hospital window with his feet if he hung down, even though the distance was
10 times his height
§ S.B. had a terrible
time with traffic when he could see, even though he hadn't previously
§ Depression very common
in these cases; often sit in the dark to regain familiar sensory environment
Neurophysiological Studies of Single Cells
§ Some of the most
important work in understanding how visual perception works has focused on the
way in which varying visual stimuli affect the firing of single cells in the
visual cortex
§ Hubel
& Wiesel's studies of feature detectors in cats
§ Perceptual features are simple characteristics of
stimulus patterns, e.g., lines, edges, line at a given angle, curve line,
movement in a given direction, movement, etc.
§ Hubel
& Wiesel recorded activity from single cells in
the brain while presenting simple visual stimuli (generally bars of light) in
front of the cat
§ Found that some cells
were only active when bar of light was presented at a certain angle. Cell would
fire long bursts of impulses at only that
angle and spot on the retina, while at other angles would show no
activity at all. Different cells would respond to different angles
§ Cells deeper in the
brain respond to more generalized characteristics such as bars at a given angle
regardless whether moving or stationary, or what part of the retina they
stimulated
Significance of Hubel & Wiesel's Findings
§ These findings are
extremely important because they show there are specific mechanisms in the
brain for selecting & detecting certain features
of objects
§ Suggests that
perceptions may be built up from combinations of these selected features
(bottom-up processing)
§ If true, then
understanding what those features are and how the brain uses them to construct
shapes, events, etc. could greatly increase our understanding of perception
Various Phenomena Showing that Perception is Modifiable by
Experience
§ Perceptual Adaptation
§ Perceptual Set
§ Perceptual
Differentiation
§ Selective Attention
Perceptual Adaptation: The World Upside Down
Experiments on the
distortion of the visual field (Stratton, Ewert,
Kohler, Taylor, Dolezal)
show that over long periods of perceptual disturbance in human subjects,
gradual behavioural and perceptual adaptation occurs
Recent Replication of Stratton's Study
Perceptual Set
§ Perceptions are
constructed from expectations based on previous experience
Perceptual Differentiation
§ Process whereby
certain distinctive features in the stimulus are singled out and attended to
§ With experience we
learn which features of stimuli are the most important ones to distinguish
among different stimuli
Selective Attention
§ Broad term to describe
various selectional effects that can easily change
from moment to moment
§ Selective listening
(cocktail party effect, dichotic listening)
§ Selective looking
(e.g., two different videotaped games superimposed; unattended game not
remembered or noticed)
Perceptual Constancies & Illusions
§ One of the central
questions in perception is the question of how we can perceive the constant properties of the distal
stimulus, given the striking variations
in the proximal stimulus
§ The fact that visual
perceptual mechanisms maintain these constancies
leads to various perceptual illusions
§ Illusions are the
result of constancies in perception
§ Three major
constancies that have received a lot of study are
§ Lightness constancy
§ Size constancy
§ Shape constancy
Lightness Constancy
§ The apparent lightness
of an object remains constant despite changes in illumination Applet
Explanation for Lightness Constancy
§ The best explanation
of the lightness constancy effect is that our perception of the lightness of an
object depends on the brightness ratio
of the object in relation to surrounding regions
§ The effective stimulus
for the brightness of any region is the ratio
of the luminance of the region to the luminance of the surrounding region
§ Since the ratio is constant, even with changes in
illumination, lightness is constant
Size Constancy
§ Perceived size of
objects remains constant whether the objects are near or far away
§ The best explanation
for how this works is that the brain does size-distance compensation
§ Depth cues are used to
estimate distance, and the learned
relation between size and distance is used to compensate for apparent size
§ Example: Linear
Perspective
§ Many illusions seem to
be the result of maintaining this size constancy compensation
Some Size Constancy Illusions
Depth Cues & Size Constancy
v Ponzo
illusion
Moon Illusion
v Moon illusion: The overhead moon
appears much smaller than the moon near the horizon, even though the sizes of
the visual images on the retina are identical
v The most common explanation is that
the brain assumes (on the basis of depth cues and previous experience) that the
horizon is much farther away than the overhead sky. Thus, the same actual size
visual image at the horizon will appear much larger than that image in the
overhead sky which is assumed to be closer
Figure Illustrating the Size-Distance Explanation
§ Regardless of its
elevation, the distance between an observer (at the center of the horizontal
line) and the moon remains constant (unfilled circles)
Shape Constancy
§ The perceived shape of
an object is independent of the angle from which it is viewed
§ The Ames Room illusion
in large part derives from shape constancy and assumptions about rectangular
corners in familiar rooms