Visual perception

Perception 1 is focused on:

perception, motion and action

1 of 259

light enters the eyes from

the illuminated environment

2 of 259

perceptions starts when

light from the illuminated environment meets our eyes

3 of 259

perception models main question

top-down or bottom-up?

4 of 259

bottom up processing assumed to be a

serial process

5 of 259

what is a serial process?

next step cannot start until current is finished

6 of 259

perception model; bottom-up? stimulus -> attention ->

perception -> thought process -> decision -> action/response

7 of 259

What is said about top-down processing

sometimes our expectations and knowledge influence cognition

8 of 259

rather than?

the stimulus itself

9 of 259

top down processing example - the letter 'B'

if we do not get the full picture, if there's letter around we think it's a 'B' if there's numbers around, we would think it's a 13

10 of 259

top-down or bottom up? questions?

is perception indirect? is perception not driven entirely by the stimulus properties?

11 of 259

additionally,

does perception depend on internal processes? do we need to 'reconstruct' in the external environment

12 of 259

constructivist approach is

top-down

13 of 259

it is the notion that perception is

the end result of a process which begins with sensory stimulation and involve interpreting the information

14 of 259

thus perception is

indirect and relies on internal processes

15 of 259

Richard Gregory's theory is good at explaining

illusions - Eye and brain: the psychology of seeing (1966)

16 of 259

Constructivist approach; sensation ->

interpretation/ inference -> perception

17 of 259

direct perception can also be called?

ecological psychology

18 of 259

it is the idea that awareness of the world (object, patterns etc) is

essentially determined by the info present to the sense

19 of 259

thus, perception is a direct process based on

sensory information

20 of 259

this is outlines in

Gibson's ''The ecological approach to visual perception''

21 of 259

James J Gibson argued for a more

comprehensive view of the visual system

22 of 259

why?

because it has been evolved to allow us to interact with the environment

23 of 259

what is the mainstream psychology view in perception?

the function of our visual system is object recognition

24 of 259

For Gibson, there is no perception without

action and no action without perception

25 of 259

Gibson's approach was/ still is

very radical

26 of 259

It's all about surfaces and

textures

27 of 259

Gibson's work started WWII when asked to train

pilots quickly

28 of 259

he was asked to filter out potential

and non-potential pilots prior to training

29 of 259

what's the most difficult part of flying?

take off and landing

30 of 259

why? to land, you need to know where you are relative to

airstrip, angle of approach, and how to modify angle so that you can aim

31 of 259

therefore, what must be important?

depth perception

32 of 259

but tests based on pictorial cues of depth didn't

work

33 of 259

Surfaces VS

plane (Gibson, 1979)

34 of 259

plane (e.g. horizontal plane) is the

abstract notion of a flat surfaces

35 of 259

it lacks some of the qualities of

a textured surface

36 of 259

such as: a surface is substantial and

is never perfectly transparent

37 of 259

a surface can also only be seen while

the plane only visualised

38 of 259

there is structure in

the surfaces that exist in our environment

39 of 259

which structures the

light that reaches the observer

40 of 259

it is not the mere stimulation by light that

leads to perception; it is the structure of light

41 of 259

Textures - Gibson; started to suggest that we shouldn't look on

depth or space perception

42 of 259

but instead?

on the perception of surfaces in the environment

43 of 259

textured surfaces are all around us (pebbles, sand, grass etc.) and

provide useful information

44 of 259

about?

distance & depth, shape & slant, layout of objects

45 of 259

how is shape & slant in everyday life?

can see stairs, this info comes from shape

46 of 259

textures should give

layouts of objects

47 of 259

optic array is the pattern of light

reaching the eye

48 of 259

why?

it is structured and contains information about the environment

49 of 259

what is a vital property of the optic ray?

it will transform as the observer moves

50 of 259

this allows information from both

the layout and shapes of objects

51 of 259

and about the

observer's movement relative to the world

52 of 259

when the person moves, the whole

perception does

53 of 259

by sitting up we can change the way

the light reaches our eyes

54 of 259

Example of transformation - the fence

each segment represents us stood at a different angle, but it is still the same fence

55 of 259

in transformation - the relative motions in the optic array correlate with

the layout of the objects

56 of 259

optic flow is the apparent

motion of stationary objects and surfaces relative to a mobile observer

57 of 259

optic flow and the focus of expansion - as proposed by

Gibson

58 of 259

Gibson proposed how we can use

optic flow:

59 of 259

he said that when we move in a straight path,

all objects & surfaces appear to expand in a radical pattern

60 of 259

the current direction of the observed is the

origin on a single point

61 of 259

the single point is called

the focus of expansion

62 of 259

during forward movements,

the focus of expansion indicates the direction of travel

63 of 259

in this way, the animal can

see where it is going

64 of 259

to change direction, the animal can

reposition the focus of expansion in that direction

65 of 259

how?

because the focus of expansion always coincides with the direction of instantaneous heading

66 of 259

flow helps us to know

where we are going

67 of 259

optic flow helps us to know

whether we are going left or right

68 of 259

What is gaze?

eye + head position

69 of 259

with mobile gaze the focus of expansion doesn't provide

information about the direction of heading since it is displaced due to eye-movements

70 of 259

optic or retinal flow? Regan and

Beverly (1982) introduced retinal flow

71 of 259

this describes the

pattern that is actually available at the retina

72 of 259

can we decompose the retinal flow pattern to

access info about our instantaneous heading through the focus of expansion

73 of 259

to ways have been proposed, 1?

use decomposition algorithms to recover an estimate of linear heading

74 of 259

2?

use the information from efferent signals already known to the systems to subtract the rotational component introduced by eye or neck movements

75 of 259

The case for mobile gaze - how many points?

4

76 of 259

1?

having mobile gaze = not necessarily detrimental to locomotion

77 of 259

2? when rotational component is

added in the optic flow field by executing eye-movements direction of heading is judged accurately

78 of 259

3? Wilkie and Wann showed

when driving in a stimulated environment, P's performed better when allowed to use natural eye-movements as opposed to visually tracking the middle of the road

79 of 259

4? Wann, et al., went further and

suggested we don't need to retrieve heading at all, we can use retinal flow to get the direction of our future path

80 of 259

flow equalization theory suggests that optic flow

is not just used for heading

81 of 259

the case with

honeybees

82 of 259

as they tend to

fly in the middle of narrow gaps

83 of 259

Srinivasan et al trained honeybees to

fly in the middle of a patterned tunnel that led to food

84 of 259

the walls of the tunnel could move - this provided

different flow speeds at each side of the tunnel

85 of 259

when motion was introduced to one of the tunnel walls

honeybees altered their flying pattern moving towards the side that appeared slow

86 of 259

Tammero & Dickinson, 2002 - works on straight paths

curvilinear trajectories are inherently asymmetrical

87 of 259

flow averaging - in a study where optic flow asymmetries on a

curved trajectory was introduced

88 of 259

instead of p's steering towards slower moving side they

averaged flow speed from the two sides to derive to a global flow speed estimates (Kountouriotis et al 2016)

89 of 259

when slow-moving vectors are clouded we perceive them to be

moving faster

90 of 259

when the faster moving points are occluded, we perceive them to be

moving slower

91 of 259

why?

as we only reply on the central, slower-moving parts

92 of 259

another use of the optic flow field in 1976 by?

Lee

93 of 259

what is this used to calculate?

the time-to-contact with an object or surface - including the strategy by gannets

94 of 259

it could divide our estimate of objects distance by our

estimate of the objects speed

95 of 259

however this info is not

readily available to use

96 of 259

Tau overcomes this since it is using the size of

the retinal imagine of the object

97 of 259

divided by

its rate of expansion

98 of 259

this means that the faster it expands -

the less time is to contact

99 of 259

time-to-contact uses the size of the retinal image of the object

divided by its rate of expansion

100 of 259

the faster it expands =

the less time there is to contact

101 of 259

Affordance - the end product of perception is not

an internal representation of the visual world

102 of 259

but the detection of

affordance

103 of 259

i.e. what does this surface or object

offer to the animal?

104 of 259

all of the potential uses (affordances) of an object are

directly perceivable

105 of 259

objects often can have

several affordances

106 of 259

current psychological state determines

behaviour

107 of 259

different species will perceive different

affordances from the same objects

108 of 259

Affordances example - Warren 1984

showed P's pictures of stairs with differently proportioned steps

109 of 259

then asked P's whether

steps were climbable or unclimbable

110 of 259

Warren 1984 - 24 P's divided into groups;

short groups (5ft4in) and tall (mean height 6ft2)

111 of 259

both groups judged stairways as

unclimbable at a riser height in proportion to their leg lengths

112 of 259

Another affordance example comes from

Will et al 2013

113 of 259

what did they do?

showed P's pics of objects with similar shape but differing in graspability

114 of 259

P's were asked to

life their arm to perform a reach-like movement

115 of 259

the onset of the muscular activity was

faster for graspable objects than non-grapsable

116 of 259

the affordance of graspability is known to

the motor system

117 of 259

summary of VP1.... infra we pick up from textures/ surfaces e.g.

WWII pilots, depth perception, distance & depth, shape & slant and layout ob objects = important

118 of 259

summary of VP1... how optic flow can be used to judge our heading & influence trajectories -

OF; during forward movement focus of expansion indicated direction of travel, to change direction reposition focus of expansion in that direction

119 of 259

summary of VP1... term affordance from a direct perception perspective -

all of potential uses (affordance) of objects = directly perceivable, diff species perceive diff affordances from the same objects

120 of 259

Perception 2:

face recogntion

121 of 259

face recognition is the most common way of

identifying people

122 of 259

face recognition differs from

other forms of object recognition

123 of 259

prosopagnosic P's unable to

recognise familiar faces

124 of 259

this can even extend to

their own face in a mirror

125 of 259

however, they have some ability to

recognise familiar objects

126 of 259

the inability to recognise faces doesn't occur because

they have forgotten the people concerned

127 of 259

as they can still recognise

voices and names

128 of 259

how many reasons have been suggested for prosopagnosia?

2

129 of 259

1?

precise descrimination

130 of 259

PD: been suggested these p's have problems in recognising faces silly because

because more precise discriminations are required to recognise different in faces than differences in objects e.g. chair and table

131 of 259

2?

specific processing mechanisms involved in face recognition

132 of 259

SPM: DeRenzi (1986) - Prosopagnostic p who was very good at making fine discriminations e.g.between Italian coins but

unable to recognise family and friends by sight

133 of 259

Ellis and Young 1988 suggested

there are face-specific processes

134 of 259

Sergent, Ohta and MacDonald (1992) - P's categorised as

living or natural VS non-living, man-made, or categories well-known faces as belonging to actors or non-actors

135 of 259

found; brain areas specifically active in face identification tended to be

forward to those active in object recognition

136 of 259

also discovered; several areas in the right hemisphere - more active in

face identification than object

137 of 259

configurational information - when we recognise a face in a photograph there are

2 major kinds of information we might use:

138 of 259

1...

information about individual features e.g. eye colour

139 of 259

2...

information about configuration or overall arrangement of the features

140 of 259

many approach to face recognition are based on

a feature approach

141 of 259

Young, Hellawell and Hay (1987) constructed faces from photography by

combing the top halves and bottom halves of different famous faces

142 of 259

when the two halves were closely aligned, P's

experienced great difficulty in naming the top halves

143 of 259

however, their performance was much better when

2 halves weren't closely aligned

144 of 259

presumably - close alignment produced

new configuration which interfered with face recognition

145 of 259

Searcy and Bartlett 1996 - reported face processing is not

solely configurational

146 of 259

and that facial distortions in photos were produced in how many different ways?

2

147 of 259

1?

configural distortions - e.g. moving the eyes up and mouth down

148 of 259

2?

component distortions - blurring pulls of the eyes to produce cataracts, blackening teeth and discolouring remaining teeth

149 of 259

the photos were then presented upright or

inverted

150 of 259

and the P's gave them

grotesqueness ratings on a 7-point scale

151 of 259

the findings suggest that

component distortions are readily detected in both upright and inverted faces

152 of 259

whereas, configurable distortions are often

not detected in inverted faces

153 of 259

thus meaning?

configurational and component processing can both be used with upright faces

154 of 259

but the processing of inverted faces is

largely limited to component processing

155 of 259

most research on face recognition has

used photos or other 2D stimuli

156 of 259

there are at least how many potential limitations of such research?

2

157 of 259

1?

viewing an actual 3-D face provides more info for the observer than does a 2-D

158 of 259

2?

people's faces are normally mobile, registering emotional states, agreement or disagreement with what is being said, and so on

159 of 259

none of these dynamic changes over time is

available in photos

160 of 259

Bruce and valentine 1988 - small illuminated

lights were spread over a face, then filmed in the dark so only lights could be seen

161 of 259

P's showed some ability to determine the sex and

identity of each face on the basis of movements of the lights

162 of 259

they were also very good at

identifying expressive movements (such as smiling or frowning)

163 of 259

Models of facial recognition - 2 major theorists

Bruce & Young 1986, Burton & Bruce 1993

164 of 259

B&Y 1986 - there are

major differences in the processing of familiar and unfamiliar faces

165 of 259

familiar faces?

primarily depends on structural encoding, face recognition units, person identity nodes, and name generation

166 of 259

unfamiliar faces?

involves structural encoding, expression analysis, facial speech analysis, and directed visual processing

167 of 259

they argued that

several different types of info that can obtained from faces and which corresponds to the 8 components of their model:

168 of 259

it consists of;

structural encoding, expression analysis, facial speech analysis, directed visual processing, face recognition units, person identity nodes, name generation, cognitive system

169 of 259

structural encoding?

produces various representations or descriptions corresponding apron to those identified within Marr's 1982 model

170 of 259

expression analysis?

Indi's emotional state can be inferred from analysis of their facial features

171 of 259

facial speech analysis?

speech perception can be facilitated by detailed observation of speakers lip movements

172 of 259

directed visual processing?

for certain purposes e.g. decide whether psychologists have beards - specific facial information may be processed selectively

173 of 259

face recognition units?

each face recognition unit contains structural information about one of the faces known to the viewer

174 of 259

person identity nodes

provide info about person concerned e.g. interests, friends, contexts in which encountered

175 of 259

name generation:

person's name stored separately from other info

176 of 259

cognitive system

contains additional info e.g. that actors/ actresses usually have attractive faces

177 of 259

cognitive systems also plays an important part in

determining which component/s of the system receive attention

178 of 259

Bruce 1988 - evidence

lab studies on norm individuals, cog neuropsychology investigations of brain damage P's and diary studies

179 of 259

malone, Morris, kay and levin 1982 - if it were possible to find

P's who show good recognition of familiar faces

180 of 259

but for recognition of

unfamiliar faces

181 of 259

another patients who showed

the opposite pattern

182 of 259

this would provide strong evidence that

the processes involved in the recognition of familiar and unfamiliar faces = different

183 of 259

so what did they do?

obtained evidence in line with these predictions

184 of 259

they tested 1 P who showed

reasonable ability to recognise photographs of famous statesmen

185 of 259

how many correct?

14/17

186 of 259

but he was severely impaired in a task matching

unfamiliar faces

187 of 259

2nd P = quite different -

performed at normal level on matching unfamiliar faces

188 of 259

but had great difficulty in recognising

faces of famous people (5/22)

189 of 259

this indicates that

the name generation component can be accessed only via the appropriate person identity node

190 of 259

the model predicts that one should never be able to

put a name to a face without at the same time having other available info about the person

191 of 259

what does this explain?

why people frequently forget names

192 of 259

young, hay and ellis 1985 - asked P's to keep

diary record of specific face recognition problems experience day-to-day

193 of 259

how many incidents?

1008 altogether

194 of 259

not once did a subject report

putting a name to a face while knowing nothing else about the person

195 of 259

there was 190 occasions where the subject could

remember a fair amount of info about the person but unable to remember name

196 of 259

most brain damaged P's who cannot put

names to faces have great difficulty in naming ordinary objects

197 of 259

in such cases as previously mentioned, it is not simply

name generation component of face-recognition system which is impaired

198 of 259

McKenna and warrington 1980 - patient

GBL

199 of 259

naming problems seemed to be specific to

faces

200 of 259

was able to accurately supply info about

90% of famous people whose photos she saw - could only name 15%

201 of 259

named 80% EU cities and

100% English town

202 of 259

according to the model, another kind of problem should be

fairly common

203 of 259

if appropriate face recognition unit is activated but the person identity node isn't, then

should be a feeling of familiarity coupled with inability to think of any relevant info about them

204 of 259

in set of incidents collected by Young et al 1985 how many occasions was this reported?

233

205 of 259

further predictions ... 1;

when we look at familiar face, familiarity info from face recognition unit should be accessed first

206 of 259

followed by

info about that person from person identity node

207 of 259

followed by

the person's nae from name generation component

208 of 259

basically... familiarity decisions about a face should be made faster than

decisions based on identity nodes

209 of 259

continued... Young et al., 1986b - discovered P's decided whether or not

face = familiar faster than whether or not it was a politicians

210 of 259

decisions based on person identity nodes should be made faster than

those based on the word generation component

211 of 259

1986a - found P's = faster to decide

whether face belonged to political than producing persons name

212 of 259

Cog neuropsychological evidence - practically no brain-damaged p's can

put names to faces without knowing anything else about the person - but several p's show the opposite pattern

213 of 259

flude et al., 1989 - Patient EST - able to retrieve occupations for

85% of very familiar people when presented with faces, but recall only 15% of names

214 of 259

overview: convincing evidence that - the model of B&Y 1986 provides

coherent account of various kinds of info about faces and ways in which these kinds of info are related to each other

215 of 259

overview: convincing evidence that - several different components =

involved in face processing

216 of 259

overview: convincing evidence that - differences in the processing of familiar and

unfamiliar faces are clearly identified

217 of 259

overview: convincing evidence that - familiar and unfamiliar faces are

typically processed quite differently

218 of 259

overview: convincing evidence that - the model proposed by B&Y 1986 is on the right lines:

a) information about familiar faces accessed sequentially and b) the order in which different kinds of info are accessed also corresponds to theoretical assumptions

219 of 259

The main inadequacies of the model relate to: 1.

insufficient specification of some of the components and processes involved in face recognition:

220 of 259

a) the cog system - B&Y 1986 serves to

catch all those aspects of processing not reflected in other components of our model

221 of 259

b) the account of the processing of unfamiliar faces is

much less detailed than the one offered for familiar faces

222 of 259

it has been found with both familiar and unfamiliar faces that

the speed and accuracy of recognition is affected by the context in which a face is presented

223 of 259

with familiar faces contextual information about individual's occupation/ where they have been previously encountered activates

person identity node - and in turn activates appropriate face recognition unit facilitating the recognition of a face as familiar

224 of 259

in the case of unfamiliar faces,

context effects have also been found

225 of 259

example - unfamiliar faces are recognised better if hey are shown for a second time against

same background context as the first showing

226 of 259

however the previous finding is

hard to incorporate within the model

227 of 259

overview: convincing evidence that - 2, evidence is inconsistent with the assumption that

names can be accessed only via relevant autobiographical info stored at the person identity node

228 of 259

amnestic P - ME could match

faces and names of 88% of famous people for whom she was unable to recall any autobiographical info

229 of 259

overview: convincing evidence that - 3, important for theory that some p's show better recognition for

familiar faces than unfamiliar faces, whereas others show the opposite pattern

230 of 259

this double dissociation was obtained by

Malone et al 1982 but has proved difficult to replicate

231 of 259

example - Young et al. 1993 - unsuccessfully studied

34 brain-damaged men

232 of 259

5 of the p's had

selective impairment of expression analysis - but there was much weaker evidence of selective impairment of familiar or unfamiliar face recogniton

233 of 259

young et al 1993 - argued that

previous research may have produced misleading conclusions because of meth limitations

234 of 259

interactive activation and competition model - Burton and Bruce (1993) developed

the bruce and young model 1986

235 of 259

one of the features is that there is separate store for names which can only be accessed

via relevant autobiographical info stored at person identity node

236 of 259

Dehaan et al., 1991 contradict this - they investigated

amnestic P - ME

237 of 259

she was able to match faces and names of

88% of famous people for whom she unable to recall autobiographical information

238 of 259

the fact her PINs were damaged should have prevented

matching names and faces

239 of 259

Burton, bruce and Johnston 1990/ Burton and bruce 1993

revised and developed bruce and young model

240 of 259

assumed there were how many pools of information?

3

241 of 259

1 - FRU

face recognition units - contains stored info about specific faces

242 of 259

2 - PINs

person identity nodes - gateways into semantic info and can be activated by verbal input about people's names as well as by facial input

243 of 259

PINs provide info about the

familiarity of Indi's based on either verbal or facial info

244 of 259

3 - SIUs and NRUs

semantic information unit and name recognition units - contain name and other info about Indi's and names

245 of 259

there are bi-directional excitatory links between

pools - etc unit it linked to others by means of inhibitory connections

246 of 259

a face is recognised as familiar when

level of activity in appropriate PIN reaches threshold level of activation; same mechanism is involved in recognition on basis of name, voice or other information

247 of 259

experimental evidence - the model has been applied to

associative priming effects that have been found with faces

248 of 259

for example; time taken to decide whether face is familiar is reduced when

face of related person is shown immediately beforehand

249 of 259

experimental evidence - according to model the first face activates

SIUs - which feedback activation the PIN of that face and related faces

250 of 259

then then speeds up

familiar decision for the second face

251 of 259

PINs can be activated by both

names and faces - follows that associative priming for familiarity decisions on faces hold be found when the name of a person e.g. prince Phillip, followed by face of a related person e.g. queen Liz

252 of 259

differences between IAC model and Bruce & Young's 1986 Model -

3 points

253 of 259

1 - no separate store for names as both stored in the

SIUs in Burton & bruce 1993 and Buce & young; name info only accessed after autobiographical info

254 of 259

2- familiarity decisions made at the

PIN level rather than FRU

255 of 259

3 - model is more

precise

256 of 259

it can account for findings of DeHaan et al 1991 - the fact that

amnesic patient ME could match names to face in spite of being unable to access autobiographical info is more consistent with Burton and Bruce 1993

257 of 259

Cohen 1990 found faces produced better

recall of names than of occupations when the names were meaningful and the occupations were meaningless

258 of 259

this couldn't happen according to the Bruce and Young 1986 model, but

proposes no problems for the Burton and Bruce 1993

259 of 259

Get Revising

Visual perception

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Similar Psychology resources:

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Related discussions on The Student Room

Similar Psychology resources: