Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

Perception & Action PSYC21012 > Weeks 3 & 4 - 'What' Pathway (Karen Lander) > Flashcards

Weeks 3 & 4 - 'What' Pathway (Karen Lander) Flashcards

1
Q

How do Humans perform object recognition?

A

Perception of familiar items between categories discrimination and within categories discrimination

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is between categories discrimination?

A

The difference between a table and a chair (belong in different object categories)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is within category discrimination (intra-class variation)?

A

The difference between 2 different chairs (they belong to the same object category)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

How do computers perform object recognition?

A

Perception of familiar patterns

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Why is object recognition difficult?

A
  • The environment contains hundreds of overlapping objects
  • Yes perceptual experience is of structured, coherent objects which we can recognise, use and usually name
  • Apparent size and shape of an object does not change despite large variations in retinal image
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is variablility in the environment which makes object recognition difficult?

A
  • There is a field of view but the object can appear in different areas
  • Can appear as differernt rotations
  • Can appear as different sizes
  • Can appear as different colours (when other tones overlap)
  • The presence of other objects can occlude the object of focus
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is viewpoint variation as a challange for recognition?

A
  • Objects can be viewed in different angles and viewpoints
  • However, we can still normally recognise the object
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is the template theory for 2D pattern matching?

A
  • We posess a mini template of a pattern (A for example) in the LTM
  • We have variations of different patterns (B for example)
  • We match the pattern to memory when we see something
  • When the match is good enough, we have recognition (greatest overlap)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What are 3 issues of the template theory for 2D pattern recognition?

A
  • Problem with imperfect matches
  • Cannot account for the flexibility and complexity of the pattern recognition system
  • Comparison requires identical orientation, size, position of template to stimuli
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What is the prototype theory of 2D pattern matching?

A
  • We do not have many different templates
  • Instead, we have a stored average template for each individual characteristic
  • We modify the templates for matching
  • No match is perfect; a criterion for matching is needed
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is Frank and Bransford’s evidence for prototype theories of 2D pattern matching?

A

METHOD
- Present ppts with visual patterns

  • These patterns were centred around a prototype
  • The prototype was not shown

RESULTS
- When asked how confident they were that they had seen different patterns, they were certain they had seen the prototype before

  • Suggests the existence of prototypes
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the feature theory for 2D pattern matching?

A
  • Patterns consist of a set of features or attributes
  • For example: the letter A has two striaght lines and a connecting crossbar
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is structural descriptions as an explanation of 2D pattern matching?

A
  • We need to describe the nature of the components of a configuration and the structural arrangement of these parts
  • For example the letter A is similar to the letter H
  • Therefore we need to know the relationship between the features rather than just the features themselves.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How is 3D object recognition similar to 2D pattern recognition?

A
  • We must interpret input to the visual system as coherent structures
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

How is 3D object recognition different from 2D pattern recognition?

A
  • We need to segregate the object from the background early in 3D
  • Must be processed to give a description which can then be matched to the descriptions of visual objects stored in memory in 3D
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are 4 encompassing questions about object recognition?

A
  1. What features are used in the structural description (primitives)
  2. How is the relationship between these elements specified?
  3. How is the overall description invariant across views (descriptions same for different people)?
  4. What about viewpoint dependence?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What is Marr and Nishihara’s cylinder explanation of how people recognise 3D objects?

A
  • Objects are comprised of cylinders and the relationship between the cyliniers is the structural description
  • Structural relations of an object are explained by a hierarchical organisation of cylinders
  • The position of each cylinder described relative to its own axis, results in a description which is invariant across viewpoints
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What is an issue with Marr and Nishihara’s cylidner approach to explaining 3D object recognition?

A
  • Doesn’t work well for everything
  • For example a scrunched piece of paper cannot be broken down into cylinders
  • We need a more flexible approach
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is Biederman’s recognition-by-components theory for 3D object recognition?

A
  • Objects are composed of 36 basic shapes
  • These shapes are called GEONS (geometrical ions)
  • We can make any object as a combination of GEONS depending on how they are organised
  • Breaking an object into a series of its constituent GEONs reveals what the structural description of the object is.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What type of theory is Bierman’s recognition-by-components theory?

A

Viewpoint invariant theory

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

What elements of object’s contour are helpful in segmenting visual images into parts as suggested by Biederman?

A
  • Concave parts
  • Areas of concavity are helpful for segregating GEONs
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What are the non-accidental properties of GEONs which aid recognition of an object?

A
  • Curvature- points on a curve
  • Parallel- Set of points in parallel
  • Co-termination- Edges terminating in a common point
  • Symmetry- versus asymmetry
  • Co-linearity- points in a straight line
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What is an example of how non-accidental properties can aid recognition of an object?

A
  • A cylinder posseses curved edges and two parallel edges connecting the curved edges
  • Regularities in the visual image are thought to reflect actual (non-accidental) regularities in the world.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What is the chain of events that occur when detecting 3D objects according to Biederman? (4)

A
  1. Edge extraction (or contour extraction)
  2. Detection of non-accidental properties and parsing of regions of concavity
  3. Determination of components
  4. Matching of components to object representations
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

What, according to Biederman, what would a lack of make object identification harder?

A

Removal of edges at areas of concavity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What is some supporting evidence for Biederman’s model where edges were deleted in different areas?

A

METHOD
- Biederman deleted edges at points where its easily reinstated or difficult to determine

  • Stimuli was presented for 100, 200 or 750 msec with 25%, 45% or 65% of contours removed

RESULTS
- When edges were removed at points of concavity, although the same amount was removed in a second image not at these points

  • There was slow and inaccurate recognition at the concave points (non-recognisable)
  • On the images where concavity remained there was relatively good recognition and the shapes were recognisable
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

What occurs when components are deleted from an image on the ability to recognise (Biederman)?

A
  • Deletion of a component affects matching stage, reducing the number of components to match to
  • Midsegment deletion also makes it more difficult to determine components
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

What are the results of the experiment conducted by Biederman on the removal of components versus midsegment elements?

A
  • At breif exposures to the stimuli (65ms) partial objects are better recognised
  • At longer exposures (200ms) midsegment deletion led to less errors
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

What is support of Biederman’s theory from Vogels et al?

A
  • They found some cortical neurons in monkeys sensitive to GEONs
  • They assessed response of individual neurons in the inferior temporal cortex to change in GEON or change in the size of an object
  • Some neurons responded more to GEON changes, providing support for GEONs
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

What are 7 issues with Biederman’s theory of 3D object recognition?

A
  • Why 36 GEONs and why a cube for example over another shape?
  • Experimental results consistent with the model but doesn’t provide critical test (may be other ways we can explain the data)
  • Doesn’t explain how descriptions are matched to those stored
  • Structure is not always key in recognition: peach vs nectarine
  • No explanation of within object discrimination
  • De-emphasises the role played by the context in recognition

-Simplifies the contribution of viewpoint dependence (easy to recognise familiar objects from one perspective over another sometimes)

31
Q

What are two advantages of Biederman’s theory on 3D object recognition?

A
  • Recognises the importance of the arrangements of the parts
  • Parsimonious: Small set of primitive shapes
32
Q

What do viewpoint invariance theoires stress?

A

That the ease of object recognition is not affected by the observer’s viewpoint

33
Q

What do viewpoint dependent theories stress?

A

They assume that changes in the viewpoint reduce the speed and/or accuracy of object recognition

34
Q

Which object representations are used predominantly: viewpoint invariance or viewpoint dependence?

A
  • Evidence suggests that viewpoint invariant mechanisms are used sometimes in object recognition whereas viewpoint dependent mechanisms are used at other times
  • Viewpoint dependent is more important for within category discriminations
  • Viewpoint invariant is more important for categorical decisions
35
Q

What is Humphreys et al’s stages of object knowledge beyond recognition?

A
  • Object
  • Structural description- about what the object is and its components
  • Semantic representation- about what we use an object for
  • Name representation
  • Name
36
Q
A
37
Q

What is an issue with Humphrey’s model of beyond object recognition?

A

There is oversimplification- as later processes may start before earlier ones have been completed (more interaction).

38
Q
A
39
Q

What is Humphrey’s cascade model?

A
  • Suggests an interaction between the structural, semantic and naming stages (between and within)
  • This makes different predictions about how subjects will perform in object naming tasks- problems at one stage will have a knock on effect
40
Q

What is evidence to support Humphrey’s cascade model?

A
  • Evidence from patients behavioural data shows there is segregation between the stages
  • However they show a cascade operation rather than independent operation
  • For example some patients know that they recognise an object but don’t know what to do with it- so cannot retrieve the semantic information.
  • Other patients know that there is an object that they can sit on (semantic information) but cannot tell you that this item is a chair.
41
Q

What is an agnosia?

A

Failure of knowledge or recognition

42
Q

What is a visual agnosia?

A
  • Feature processing and memory remain intact, and recognition defecits are limited to the visual modality
  • Altertness, attention, intelligence and language are unaffected
  • Other sensory modalities (touch, smell) may substitute for vision in allowing objects to be recognised.
43
Q

What is an apperceptive agnosia?

A
  • Problems with early processing (shape extraction)
  • Perceptual defecit, affects visual representations directly, components of visual percept are picked up but can’t be integrated, effects may be graded, often affected: unusual views of objects
44
Q

What is associative agnosia?

A
  • Problems with later processing (recognition)
  • Visual representations are intact, but cannot be accessed or used on recognition. Lack of information about the percept
  • Normal percepts stripped of their meaning
45
Q

What are the demands of face recognition?

A
  • Recognition in context
  • Object invariance
  • Specificity (IMPORTANT)
46
Q

What is the results of 1970 research into eye witness testimonies on how well we store facial knowledge?

A
  • A man named Lazlo Virag supposedly stole money and held a police officer at gunpoint
  • There were many witnesses to the situation
  • One witness even stated that his face was ‘imprinted’ into thier brain
  • The man was arrested
  • It later turned out it was not this man who commited the crime
  • Therefore, should we be relying on eye-witness testimonies to identify people?
47
Q

What is a study by Ellis into the effectiveness of the photofit system?

A

METHOD
- Two groups
- 1, were given a photo of a person and had to do a sketch and a photofit of this person
- 2, were shown a photo of a person (then was taken away) and had to do a sketch and a photofit of this person

  • The mean likeness (max 100) of the witnesses own sketches and photofit were assessed.

RESULTS
- When the picture was present, sketching was significantly better than the photofit
- When the picture was absent, photofit and sketch likeness were similar with photofit slightly better.

CONCLUSION
- Photofit is not a very good way of forming representations of the face and people have difficulty reproducing the likeness of even familiar faces

48
Q

How does photofitting work?

A

People are given a list of features in which they need to pick out individual features. The system then maps this onto a face and accumulates the different features together.

49
Q

What does configurally mean?

A

As a pattern or holistically

50
Q

What does it mean when it is thought that faces are represented configurally? (3 items)

A
  • The spatial relationships between features are as important as features themselves
  • That face features interact with one another
  • That faces are processed holistically
51
Q

What did Bradshaw and Wallace find about features versus configurations?

A
  • Participants are presented with two faces and they have to determine whether they are the same or different
  • The more different the faces are, the easier it is to distinguish
52
Q

What is a study done by Tanaka and Farah into configurations and features effecting memory of a face?

A

METHOD
- There are 2 groups
- 1 see a picture of a normal face and learn this face as ‘Larry’
- 2 see a picture of a scrambled face and learn that this is ‘Larry’

  • The participant is then asked to pick out which is Larry’s nose (2 options)
  • Also have to pick out which is Larry’s face (2 options)

RESULTS
- Group 1 who learnt the normal face were better at picking out between the two faces instead of the feature
- Group 2 who learnt the scrambled face was better at identifying the single feature instead of the whole face configuration

CONCLUSION
- This is evidence that we remember faces as an overall pattern (find it hard to determine individual features).

53
Q

What is a study conducted by Young et al onto composite versus non-composite faces?

A

METHOD
- Participants are given an image of a face where the top half is one celebrity and the bottom half is a different celebrity but the images are either 1. Aligned (composite) or 2. Misaligned (non-composite)

  • They were asked to name as quickly as possible the individual celebrities and reaction time was measured

RESULTS
- Participants were significantly slower to name composite than non-composite stimuli

CONCLUSION
- Percption of a novel facial configuration interfered with the identification of the constituent parts

54
Q

What is a study conducted by young et al onto composite/non-composite upright versus inverted faces?

A

METHOD
- Conducted the composite and non-composite study where different groups had either an aligned image of two celebrities or misaligned

  • This time however, they had one of the previous conditions and either the image was the right way up or inversed
  • Reaction time on identifying the celebrities was still measured

RESULTS
- when the image was inversed, there was no difference in reaction time between composite and non-composite images

  • Unlike normal way up where reaction time was faster for non-composite images

CONCLUSION
- Interference from the configurational infomation was only found for upright faces

55
Q

What is Thompson’s Thatcher illusion study?

A

METHOD
- Participants were given either an image of an upside down Margret Thatcher where the mouth and the eyes were turned the other way
or
- Participants were given a right way up MT whose mouth and eyes had been turned the other way

  • Participants were asked to rate the images of MT on their degree of ‘grotesqueness’

RESULTS
- Participants rated the upside-down MT as being less grotesque than the right way up one

CONCLUSION
- When we process the upright framework of the face, we process as a holistic pattern and as it does not conform to the pattern it appears grotesque

  • When the image is upside down we process as features so does not matter so much (lack of configural processing for inverted faces)
56
Q

What are 2 examples of surface properties (primal sketch) needed for face identification?

A

Pigmentation

Shading

57
Q

What is Bruce and Langton’s study on the effects of negation (removing surface characteristics)?

A

METHOD
- Participants are given a black and white image of a famous person in which they have to identify

  • They then have to identify it upside down
  • Then a negated image (invert the colours)
  • Then an upside down negated image

RESULTS
- There was lower recognition of the celebrity when the image was upside down

  • There was even less recognition for the celebrity when it was negated
  • Maximal lack of recognition occured when the image was upside down and negated

CONCLUSION
- Negating the image disrupts the 3D structure of the face and reduces the ability to recognise

58
Q

What are two possible sources of information adversely affected by negation? (Facial Processing)

A

Surface Characteristics

  • Pigmentation- skin colour and hair colour variation
  • Pattern of shading and shadow which may help specify 3D structure of a face
59
Q

What happens when an image that lacks pigmentation (a 3D head scan) is negated?

A

The effect of negation is reduced
The initial image only contains 3D structure, when we negate this image it therefore does not have an effect

60
Q

How did Bruce and Hanna et al study the effects of edges and texture on face recognition?

A

METHOD
- Participants are presented with one of three images of a celebrity

  • 1 only has edge information
  • 2 only has texture information
  • 3 was the actual picture

RESULTS
- Edge and texture information lead to less recall but texture had a higher percent than edges.
- This shows the importance of texture information in facial recognition.

61
Q

How did Valentine and Bruce study the effects of distinctiveness and caricature on facial memory?

A

METHOD
- Participants are presented with images of famous and unfamiliar faces and asked to rate ‘how well it would stand out in a crowd’

  • This allowed the establishement between typical and distinctive faces
  • There were two tasks:
  1. Participants are asked if the person they are presented an image of is familiar or not and reaction time is measured
  2. Participants are asked whether the image they are being presented with is a face or not

RESULTS
- In the 1st familiarity task, reaction time was faster when the face was more distinctive

  • In the 2nd categorisation task, reaction time was faster when the face was more typical
62
Q

What is Valentines face space scale and what does it show?

A
  • It presentes an axis where the x is a scale of a feature for example large eyes on one end and small eyes on the other end
  • The y axis is another feature for example dark hair and light hair
  • People are located on different areas of the plane depending on their eye size and hair colour
  • Nearer to zero is the most clustered ‘typical faces’
  • Spread out towards the extreme ends are the outliers with more ‘distinctive faces’
  • It shows that when there are no other faces near a distinctive one on a plane- it will be easier to identify this person. But at the centre of the space where the typical faces are would be easier to tell if it is a face or not as there are many examples of typical faces.
63
Q

What are the 4 stages in recognition of familiar faces in Bruce and Young’s model?

A
  1. Form a structural description of seen face (structural encoding)
  2. Match to stored representations of known faces- Face Recognition Units (FRUs). Each FRU contains structural information about a face known to the viewer
  3. Access semantic information- Person Identity Nodes (PINs)
  4. Recall a person’s name- name retrieval
64
Q

What are 4 supporting bits of experimental evidence for Bruce and Young’s model?

A
  1. Naturally occurring errors of face recognition where the earlier on a person gets stuck, can no longer progress onto the next stage
  2. Tip-of-the-tongue state. Giving cues about a previous state do not help, but partial information from the stage we are stuck at does.
  3. The further down the stages we go, the longer it takes to complete the task even when controlling for the more difficult stages
  4. Neuropsychological patients with different patterns of face recognition difficulty
65
Q

What other processes occur in Bruce and Young’s model alongside facial recognition?

A
  • Expression analysis
  • Facial speech analysis
  • directed visual processing

(all modulated by the congitive system)

66
Q

What is the supporting evidence for Bruce and Young’s theory of independent processing of expressions and visual speech in parallel to identity?

A
  • We can recogise expression regardless of identity
  • Propospagnosics often can recognise expressions but not identity (and vice versa)
67
Q

What are the two main issues with Bruce and Young’s model?

A
  • Independence
  • Covert recognition
68
Q

How was it found that identity and expressions weren’t as independent as Bruce and Young proposed?

A
  • Was found that reaction times for identity judgements were independent of variations in expression and facial speech

BUT

  • Reaction times for expression and facial speech were influenced by variations in identity

(for example reaction time increases for identifying someone smiling if we know them)

69
Q

How does the presence of covert recognition weaken Bruce and Young’s model?

A
  • Was found that some prosopagnosic patients with no overt recognition of famous faces showed covert responses
  • This suggests that there is leakage from the rest of the cognitive system in face recgonition

OR

  • There are two face processing routes- one conscious and one unconscious
70
Q

What did patient PH show about covert recognition?

A
  • PH was prosopagnosic but there was evidence for preserved covert recognition
  • PH was better at deciding that two views of a famous face belonged to the same person than two pictures of an unfamiliar face
  • PH showed interference from distracting faces when asked to classift names into semantic categories
  • PH showed faster recognition of familiar faces if the face or name of a related person had just been seen. Priming from famous face (doesn’t recognise this person either)
  • PH could learn true names and occupations of famous faces significantly faster than untrue ones even for faces since the accident
71
Q

What model by Burton, Bruce and Johnston can account for covert recognition?

A

connectionist ‘interactive activation’ model (IAC)

72
Q

What does the IAC model suggest?

A

Pools of units- feature units/ FRUs/ NRUs/ PINs/ SIUs are connected by bidirectional excitatory links

Activation flows around the system

73
Q

What’s an example of how the IAC model workd?

A
  • See a face (Prince Willium FRU)
  • Activates the feature units (William PIN) and passes information on
  • From seeing this one person, all related people will have activation (royals)
  • This categorical activation is passed through the system
  • Then, when we see an image of a person within the same cateogory (Kate PIN) there is feedback from the PINs and we can recognise this person faster (Kate FRU)
  • The semantic activation had been passed onwards

Perception & Action PSYC21012 (4 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions