Brainscape's Knowledge GenomeTM

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


See full index

Visual Neuroscience > W11 - Binocular Depth Perception > Flashcards

W11 - Binocular Depth Perception Flashcards

1
Q

What are the 2 assumptions that the Marr Poggio stereo model assumes?

A
  1. objects tend to be opaque, thus along a given line of sight, there can only be one match, ie. inhibitory cells activated along a line of sight
  2. objects tend to be bigger than a receptive field size in V1, and also objects tend to be SMOOTH, not jagged, ie. alone a depth plane, neighbouring cells will have the same depth match, ie. excitatory response occurs along the same depth line
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is vergence and how can it be used as a depth cue?

A

Vergence occurs when you are looking at less than 6m, the eyes move in OPPOSITE directions to focus on objects at different depths / linked to accommodation

  1. Convergence = When an object MOVES TOWARDS YOU, the eyes CONVERGE (rounder accommodation)
  2. Divergence = when an object moves away from you, the eyes DIVERGE (flatter accommodation)

Vergence could be an absolute cue for depth

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

What are the 2 types of binocular cues?

A
  1. Vergence (can be convergence or divergence)
  2. Binocular disparity
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

How do you test for convergence as an absolute cue in humans?

/

What experiment was conducted to show that we use it as a depth cue? (DICHOPTICALLY PRESENTING DISKS)

A
  1. DICHOPTICALLY present a different image to the two eyes with luminance discs
  2. Vary the lateral separation between the discs to change the required amount of CONVERGENCE
  3. Measure participant’s perceived size of the disk with varying convergence
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

If convergence IS used as a cue to depth (IT IS) - how should the perceived size of the disk vary as convergence is increased ?

A

1.Considering how further perceived distances to objects are perceived as larger,
2. You will UPSCALE the retinal image size of the disc by a GREATER amount if you DIVERGE compared to CONVERGE for the same size discs,
Increasing convergence results in a smaller estimate of viewing distance by the visual system, resulting it smaller perceived size - even if physical size of disks remain the same

  1. THUS if vergence is being used as a cue to distance, perceived size will be dependent upon the VERGENCE STAGE OF THE EYES
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

In terms of location, what are the two main ways that the eyes can be placed?

A

There are 2 main ways the eyes and thus monocular visual fields can be arranged:

  1. Overlapping = this reduces the total visual field that the person can see
  2. Non-overlapping
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the functional consequences of these two different placements?

+ Hence what type of animals would have the different placement locations?

A
  1. Functional consequences
    • Greater binocular vision gives us greater depth perception
    • greater monocular field of view allows for greater spatial or temporal acuity over a wider field of vision
  2. Animals = humans have greater binocular vision for enhanced depth perception while other birds and rabbits that seek prey have mostly monocular field of vision for wider visual coverage to catch prey
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is decussation and what is the functional consequence of it?

A

the TOTAL the crossing over of optical nerve fibres from one hemisphere of the brain to the other

Functional consequence = the ratio of crossed to uncrossed fibres is proportional to size of the binocular visual field

eg. rabbits with decussation have 100% crossing over, very little binocular overlap

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is hemidecussation and what is the functional consequence of it?

A

the partial crossover of optical nerve fibres from one hemisphere to the other - (occurs in some mammals and humans that have significant binocular overlap)

Functional consequence = the ratio of crossed to uncrossed fibres is proportional to size of the binocular visual field

In humans, 50% have uncrossed fibres, 50% of fibres of the left eye cross over to the right hemisphere and vise versa

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What would happen if you didn’t have hemidecussation?

A

Without hemidecussation, there would be a discontinuity in the SPATIAL TOPOGRAPHIC MAPPING IN THE V1 VISUAL CORTEX and you wouldn’t be able to perceive objects as a coherent percept in the brain

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is binocular disparity (also called retinal disparity) and why does it occur?

A

Binocular disparity is the offset between non-corresponding points from both eyes, which forms the basis of binocular depth vision / stereopsis

The actual offset of the image from the corresponding point

it is caused by interpupillary distance between the two eyes, causing images of objects to fall on the same and slightly different locations

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is meant by corresponding points?

A
  1. the relative location of object falls on the same location in both eyes, normally whatever both eyes are fixating on
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is zero binocular disparity?

A

the relative location of the image of the left and right eye fall exactly on corresponding points

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Depth Coding: What aspects of binocular disparity are used to encode depth?

A
  1. disparity direction = is the object left or right of the corresponding point? eg. is the object near or far relative to fixation?
  2. disparity magnitude = (how far)
    does the object fall from the corresponding point

The magnitude and the direction of the binocular disparity linked to the objects depth relative to the fixation point are used to encode depth

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the horopter and why is it important with respect to using binocular disparity to determine depth?

A
  1. The horopter is the surface on the eye that produces zero retinal disparity centered on the fixation point that moves with the eye

thus the pattern of retinal disparity needs to be interpreted in line with the shape of the horopter

  • images falling on the horopter generate zero retinal disparity
  1. The pattern of retinal disparity is always relative to the shape of the horopter produced by different viewing distances
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

How does the shape of the shape of horopter vary with viewing distance?

A
  1. At short distance = the shape of the horopter becomes concave: changes the magnitude of retinal disparity because the surface is away from fixation and getting further away from the horopter
  2. At a few metres, (abathic distance) horopoter is shaped as flat/fronto parallel plane = retinal disparity of the wall will be uniform 0 across the entire image
  3. At the longer distance, shape of horopter is convex = the magnitude is retinal disparity is decreasing as the surface is falling in front of horopter
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

In terms of determining depth using binocular disparity, why do you need to know more than just the pattern of binocular disparity?

A

The brain needs to know the shape of the horopter to be able to interpret the pattern of retinal disparity

as the shape of the horopter changes as a function of viewing distance, such that it becomes concave in close viewing distance, flat (abathic) at a distance of a few metres and convex at longer distances

The horopter is the surface that produces zero retinal disparity, and thus the pattern of retinal disparity needs to be interpreted in line with the shape of the horopter as RD is relative to horopter shape

18
Q

What pattern of binocular disparity would be produced by a flat surface at the
following distances?
a. Closer than the abathic distance
b. Abathic distance
c. Further than the abathic distance

A
  1. At concave distance = point of fixation will be 0, other non-corresponding points of the wall fall BEHIND the horopter, pattern of retinal disparity shows increasing far retinal disparity
  2. At abathic distance = pattern of retinal disparity of all points of the wall will be uniform 0
  3. At convex distance = pattern of increasing retinal disparity but surface falls IN FRONT of horopter
19
Q

What is Panum’s fusional area?

A

the distance and depth around the horopter with which you can see a single fused image (region in visual space over which we perceive single vision)

20
Q

What is a stereogram and how do they work?

A

A stereogram is a two-dimensional image that creates the illusion of depth and appears three-dimensional when viewed in a specific way

  1. By lacing two nearly identical images side by side. When viewed by focusing behind the image / diverge (left eyes goes more left and vise versa) or fixate on your finger in front of the image and look beyond such that the left eye looks at the right side of the image the brain interprets the newly focused level as the level of the image, causing depth perception.
21
Q

What is a telestereoscope and why are they used? / eg. NASA

A
  1. a binocular optical instrument used for stereoscopic viewing of distant objects; a small range finder
  2. Telestereoscopes are used to increase magnitude of perceive depth and retinal binocular disparity by increasing the horizontal offset / distance between the two images
  3. This is done by NASA to spacecraft taking pictures of the sun - taking really detailed 3D photos with STEREO Solar Terrestrial Relations Observatory, takes a slightly different images, comparing the 2 images to generate 3D detailed images of the sun
22
Q

What are the six types of stereograms?

M - A - P - P - S - A

A
  1. Mirror stereograms 1838
  2. Prism stereoscope
  3. Anaglyphs (blue/red 3D glasses)
  4. Polaroid stereoscope
  5. Shutter stereoscope
  6. Autostereogram
23
Q

In stereopsis, what is the correspondence problem?

A

The brain’s task of determining which images in the eyes should be matched and combined because they correspond to the same object out in space
This is judged on an object by object basis, as objects are at different depths and the disparity offset differs:

24
Q

What is a Keplerian Projection?

A

shows all possible binocular matches for N objects = N^2 possible matches

25
Q

If there are N objects in the field of view, how many potential matches can the visual system make, and hence, how many of them are false matches?

A
  1. Potential matches = N^2
  2. False matches = N^2 - N

such that if there were 4 object, 4^2 = 16 possible matches, 16-4 = 12 false matches

26
Q

What experiment was conducted to show that the visual system does NOT use monocular features to solve the correspondence problem?

A

(Julesz) = tested whether monocular cues were used to solve correspondence problem by removing form information

  1. Method = random-dot stereogram, using 2 pictures of random fields of black and white dots.
  2. Cut out a region and more it to the left or to the right, when stereoscopically fused, a centre square is seen floating above the background in vivid depth
  3. Thus you cannot perceive depth with only monocular form information, depth information only appears after binocular fusion of images

“Region in depth cannot be identified by its form prior to binocular matching, and its form can only be identified following binocular matching and subsequent depth extraction”

27
Q

What stereogram can also test stereoblindness?

A

a random dot stereogram

28
Q

What are the steps involved in making a random-dot stereogram (RDS)?

A
  1. Method = random-dot stereogram, using 2 pictures of random fields of black and white dots.
  2. Cut out a region and more it to the left or to the right, when stereoscopically fused, a centre square is seen floating above the background in vivid depth
29
Q

The cells that form the neural basis of stereopsis need what three properties?

A
  1. binocular cells receiving input from both eyes
  2. monocular receptive fields in similar locations (including cells that have a range of offsets to their relative location of the receptive fields going in both directions with various magnitudes) (tuned to different retinal disparities)
  3. network that responds to patterns of inhibition and excitation
30
Q

What is the issue with the theory that monocular object recognition happens first?

A

makes object processing much more difficult, from V1 onwards the cells are all binocular

Solution= you do binocular combination and depth first, and use differences in the depth of objects to assist in object recognition

31
Q

In solving the correspondence problem, what is a false match and what is the functional consequence of it?

A

False match occurs when each image being matched in each eye do not correspond to the same object,

  1. Functional consequence = the relative location of images in 2 eyes corresponds to an object at a particular depth when there is such object there / illusion
32
Q

What approach does the Marr-Poggio use to solve the false-matching problem?

A

1, Start with an inverse-Keplerian projection: a network of cells that are encoding things at different depth locations - laying out all possible stereopsis matches

  1. First, use similarity rule = map out receptive field, it only flags a potential match if the luminance polarity of the two receptive fields of the left and right eye are the same
  2. Apples the opacity and continuity constraints via E and I interactions between cells.
    * Implements assumptions
    = inhibitory cells encoding the same line of sight
    = excitatory cells encode the same depth plane
  3. A stable solution emerges that is compatible with disparity information, opacity and continuity constraints.
33
Q

In relation to the Marr-Poggio model, what is the inverse-projection/disparity
matrix?

A

inverse-Keplerian projection: a network of cells that are encoding things at different depth locations - laying out all POSSIBLE stereopsis matches

34
Q

What two ecological constraints are used to determine the type of connections between cells in the Marr-Poggio model and what are those connections?

A
  1. The opacity and continuity constraints
  2. The opacity constraint refers to the assumption of the visual system to assume objects are opaque, leading to inhibitory response of V1 cells along a given line of sight
  3. Continuity constraints refers to the assumption of the visual system that objects in the real world tend to be bigger than the size of V1 receptive fields, leading to excitatory response of V1 cells along the same depth plane, as their assume neighbouring cells will be encoding the same depth plane
35
Q

Explain the three steps used in the Marr-Poggio model to solve the
correspondence problem

A

1, Start with an inverse-Keplerian projection: a network of cells that are encoding things at different depth locations - laying out all possible stereopsis matches

  1. First, use similarity rule = map out receptive field, it only flags a potential match if the luminance polarity of the two receptive fields of the left and right eye are the same
  2. Apples the opacity and continuity constraints via E and I interactions between cells.
    * Implements assumptions
    = inhibitory cells encoding the same line of sight
    = excitatory cells encode the same depth plane
  3. A stable solution emerges that is compatible with disparity information, opacity and continuity constraints.
36
Q

. Explain why in the Marr-Poggio model the correct solution typically evolves?

A
  1. process of our perception of depth evolves over time, indicative of a neural network model with impoverished information slowly coming up with a percept over time, first apples keplerian projection, similarity rule, then encodes information at patterns of E and I
37
Q

How can the manner in which the Marr-Poggio model encodes assumptions about the world explain the fact that we see the Ames room illusion, when
viewed monocularly, even when we know what the actual shape of it is?

A

“Marr Poggio Model is relevant to the Ames room illusion because both make assumptions at the outside that are NOT represented at a cognitive level but rather encoded by patterns of excitation and inhibition, and thus cognitive awareness of an illusion in the outside world will not reduce your susceptibility to the illusion”

38
Q

What is aliasing in terms of depth coding?

A
  1. Aliasing in depth perception occurs when the disparity offset is above 180 degrees (lower than 360 degrees) results in ambiguous depth
39
Q

What role does SF play in depth coding in relation to aliasing, and the magnitude of the depth offset that can be encoded?

A
  1. In aliasing = A low SF can encode a large depth offset quicker than a high SF before it aliases, aligns with coarse-to-fine model that we process coarse information before finer details
  2. In regard to disparity magnitude, LARGE depth offsets are best encoded with LOW SFs
    Smaller depth offsets are best encoded by high SFs
40
Q

In relation to depth processing, what is meant by coarse-to-fine processing, and what are the consequences of it in terms of the magnitude of the depth offset
that can be processed?

A
  1. Coarse-to-fine processing = We match low SF first / large depth offset, then higher SF with finer depth
  2. In regard to disparity magnitude, LARGE depth offsets are best encoded with LOW SFs
    Smaller depth offsets are best encoded by high SFs
41
Q

Does the visual system only match specific or similar SFs? How can the Random Dot Stereograms be used to test this?

A
  1. Use 2 images with different SFs with filtering to make it consistently low or high SFs (low or high pass filter)
  2. Contrast the left eye to the low SF image and the right eye to the high SF image - can you still see depth in it?
  3. The eyes can handle some visual differences in SF, but has to be an overlap to perceive depth

Visual Neuroscience (6 decks)

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