Seeing and hearing lecture 4

?
What does the retina encode?
It starts to encode relative brightness and contrast/deifference
1 of 52
What does organisation does the LGN do?
Organises Marvo/parvo cell information from the two eyes
2 of 52
What is the visual system about?
Extracting visual information rather than just representing every "pixel", for example as we move up the visual cortex things get more and more specific
3 of 52
What are the gestalt laws of perceptual organisations?
Rules for how we go from patterns to objects
4 of 52
What is the law of pragnanz?
Everytime we look at something there are multiple ways in which we can see things but we tend to see the same thing the same way all the time and agree on it
5 of 52
What is the Gestalt law of perceptual organisation group features?
We readily group visual features into whole objects
6 of 52
What is Gestalt about?
The whole is greater than the sum of the parts
7 of 52
What is the Gestalt law of proximity?
We group things which are close together and see them as belonging together
8 of 52
What is the Gestalt law of similarity?
We group things which are similar together
9 of 52
What is the Gestalt law of good continuation?
We tend to prefer the grouping where there is not a big change in direction, we assume contours continue on a smooth path
10 of 52
What is the Gestalt law of closure?
We assume things continue and are a complete object/shape
11 of 52
What explains the gestalt laws?
Things in the natural world tend to behave in the same way. things tend to be close together and lines tend to continue
12 of 52
What is figure-ground segregation?
The process by which we decide what is the object and what is the background.
13 of 52
What is theorised about figure-ground segregation?
What we see as the figure affects depth and speed of processing, we are quicker and more accurate to notice a line if it appears on the thing you are seeing as the figure.
14 of 52
How do we decide what is the figure and what is the ground in figure-ground segregation?
1) The smaller region is the figure 2) We assume the most symmetrical region is the figure 3) we assume regions below the horizon are more likely to be perceived as the figure 4) convex (bow in or out) regions are more likely to be the figure
15 of 52
What research was done to see when we decide what is the figure and what is the ground?
Vecera flashed people figure/ground pictures and found that participants are faster to recognise the figure region. Also if an attentional cue was flashed on one side of the shape you were more likely to see that as the figure.
16 of 52
When do we decide what is the figure and what is the ground?
Before we pay attention to something
17 of 52
What are the issues with the Gestalt laws?
may be limited to two-dimensional patterns it is not clear how they are applied to real scenes or 3D , there is also no explanation as to why it works like this or how the brain is doing this.
18 of 52
How good are we at recognising objects?
Very good, can recognise things in lots of different environments
19 of 52
Are computers good at recognising objects?
No
20 of 52
What is Marr's computational theory of object recognition?
It outlines the representations and algorithms necessary for vision. Primal sketch-> 2 1/2 D sketch -> 3-D model representation -> Semantic system
21 of 52
According to Marr what is the goal of object recognition?
To derive an object centred model of what we are looking at. (The thing which is coming in is a 2D image but it is not that useful for recognising objects, we ant a 3D model of what we are looking at
22 of 52
What is a viewer centred view in Marr's computational theory of object recognition
Something we see from our point of view
23 of 52
What is the imput image in Marr's computational theory of object recognition?
The first image about light and dark
24 of 52
What is the Primal Sketch in Marr's computational theory of object recognition?
Looking at lines and edges
25 of 52
What is the 2 1/2 D sketch in Marr's computational theory of object recognition?
The layers of the different layers and shapes, looking at depths,
26 of 52
How can edges be identified in Marr's computational theory of object recognition?
By using the second derivative (zero crossings) of variation in intensity over space (we look at the difference between light and dark and mark the edge where it changes)
27 of 52
What is the principle axis?
When we detect the boundaries of objects and then we work out where the cylinders are
28 of 52
What is Biederman's recognition by components model?
We detect the "non-accidental properties" in an image and work out where the geons are (3D shapes)
29 of 52
According to Biederman how many geons are there?
36
30 of 52
When are objects easier to recognise according to Biderman?
If their concavities are preserved (concave bits)
31 of 52
What evidence did Biederman and Cooper find for recognition-by-components?
The used reptition as priming as evidence for geons, objects are named as fast as possible. Repeated objects are named faster. Shown a different half of that object or a different example of that object. Quicker when we see contours, did it for geons
32 of 52
What evidence did Biederman and Cooper find for recognition-by-components using geons?
Showed people the complement geons or same name different exmplar and found that visual priming only occurs when geons are repeated
33 of 52
What are viewpoint-independent models?
That the observer's goal is to derive an object-centred representation independent of viewpoint (Marr (computational theory of object recognition) and Biederman (recognition by components))
34 of 52
What are viewpoint-dependent models?
Human object recognition is actually very dependent on view point (Tarr, Bulthoff)
35 of 52
What is evidence for viewpoint-dependent object recongition models?
Errors and reaction time to recognise objects is related to the degree of rotation, if people see objects how they normally do they recognise them faster.
36 of 52
What are issues with structural models? (where we go from one step to another)
The emphasis bottom-up processes but ignore conceptually-driven recognition. They may account for "basic-level" categorisation but not subtle distinctions. Some objects don't have geons like clouds.
37 of 52
What is visual agnosia?
A specific problem with recognising objects, there is evidence that object recognition requires special mechanisms which can go wrong. A profound, acquired deficit in recognition of objects by sight
38 of 52
What is the object recognition problem in visual agnosia NOT down to?
Visual acuity, visual field loss or colour vision (nothing wrong with their sight!), Not an issue with object knowledge as they can describe objects from memory. Not a problem with naming as they can name objects when touching them.
39 of 52
How do we diagnose visual agnosia?
It is diagnosed as a failure to name line drawings (pictures)
40 of 52
What is apperceptive visual agnosia?
Deficit in perceptual processing, cannot copy an object, have an issue with perceiving objects, issues with putting lines into shapes
41 of 52
What is associative visual agnosia?
Deficit in linking percept with semantic knowledge, can copy pictures but cannot name them , can perceive objects fine but do not have the ability to name them.
42 of 52
What is integrative agnosia?
Could copy well but could not recognise objects, said the nperson could not integrate object features into a recognisable form. When they copied they did a tiny bit at a time like they could only do feature by feature.
43 of 52
What is scene perception?
Despite their complexity, scene categories ("gist") can be recognised very quickly
44 of 52
What is Rapid Serial Visual Presentation? (RSVP)
When lots of images are shown at once
45 of 52
What does Rapid Serial Visual Presentation? (RSVP) show?
People are quick at recognising the images shows people get the "gist" of a picture quickly, quick at recognising scenes even if we don't know the objects inside them
46 of 52
What is change blindness?
When something changes between two images and you fail to recognise the change.
47 of 52
What is a limit of scene perception?
Change blindness, boundary extension
48 of 52
What is boundary extension?
After seeing a scene we tend to extend the boundaries of a picture and add in other detail which wasn't in the original picture
49 of 52
Does scene context helps object recognition?
Yes
50 of 52
What is the experiment Palmer did with scene context and object recognition?
Flashed a contextual scene and then an object and got people to try and recognise it. Sometimes the object was appropriate for the scene, one where it was inappropriate but similar and one where it was inappropriate and dissimilar.
51 of 52
What were the results form the experiment Palmer did with scene context and object recognition?
People were better at recognising the appropriate object and worst at recognising inappropriate similar because the scene confused them
52 of 52

Other cards in this set

Card 2

Front

What does organisation does the LGN do?

Back

Organises Marvo/parvo cell information from the two eyes

Card 3

Front

What is the visual system about?

Back

Preview of the front of card 3

Card 4

Front

What are the gestalt laws of perceptual organisations?

Back

Preview of the front of card 4

Card 5

Front

What is the law of pragnanz?

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Visual System resources »