The flashcards below were created by user
on FreezingBlue Flashcards.
Clues that the brain uses to distinguish depth of objects.
7 Monocular clues:
- aerial perspective
- geometric (linear) perspective
- Size (SILO, familiarity, looming)
- Texture gradient
- Motion (motion parallax, relative velocity, looming)
Closer objects partially block the view of more distanct objects. Related to perceptual phenomena such as transparency and illusory contours.
Example of illusory contours
Stimuli having lower luminance contrast with respect to the background are perchieved as farther away. Less distinct borders
When you percieve difference in depth due to differences in color.
Perception of shape and depth can be affected by how a scene is illuminated.
shading and shadows
parallel lines extending off into the distance in real world, 3D space project as converging lines on the retina. the retina cannot have a parrellel image on it bc it is curved.
Distance is indicated on a flat picture by convergence to the vanishing point.
The point where all of the lines converge, may or may not be in the scene.
When two lines are drawn in a picture with converging lines and look as though they are different length when they are actually the same size
2 objects moving across the visual field with the same linear velocity, but the one with a higher angular velocity (closer to you) appears to move faster.
(it takes less time for the closer object to cross the retina so it appears faster)
Example of relative velocity
retinoscopy, nearing the neutral point
While moving one's self and fixating on a point in space.
Motion parallax: more distant objects move more than the point of fixation; wil appear to move in the...
same direction as your head
The point of fixation remains motionless (zero velocity) and the objects that are more distant than point of fixation appear to move faster
Velocity of motion
The size of an object's image on the retina provides _____ info regarding the object's size and distance from the viewer.
- (a small retinal image may equally represent a small, near object or a large distant one)
An object that gradually increases in size in all dimensions at the same rate is perceived as moving towards observer.
Types of texture perspective:
- density scaling
- aspect ratio
This indicated inclination only if the true shape of the texture element is known of an image of a single texture.
This changes the aspect ratio of the image as a whole, as well as of the individual texture elements
- (ex: checkerboard at an incline)
You know that a quater is a circle, even if turned at different perspectives because of
The pervcieved color of objects remains relatively constant under varying illumination conditions.
Ex: you know what a rubiks cube looks like, so you can guess the color under different illuminations
Known objects are perceived as having relatively constant size despite the fact that their retinal images may vary greatly in size.
Ratios of size and position within an image provade a ____ by which size co ntancy may be obained.
An object's perceived size depends upon its perceived distance fom the viewer.
For two objects to have the same retinal image size the closer one must be smaller and the more distant one must be larger
SILO (small in, large out)
Examples of SILO
BO prism test & "over minusing"
Room that is distorted in shape in such a way as to appear normal chen viewed monocluarly from a specfic point at which normal 3D perspective can be achieved.
The Ames (Trapezoidal) room
The actual distance from the moon vs what the percieved distance is, both arcs drawn. We percieve the moon as farther away from us when it is at horizon (subtends a smaller angle on our retina than it does when it is overhead)
The importance of cues changes with distance
Depth perception theory
- *aerial cues only work in distance and convergence only works up close
- graph in notes (depth contrast, depth in meters)
Erroneous perceptions due to incomplete, ambiguous or contradictory visual info.
Illusions that involve ambiguous or misleading monocular depth cues
The lines between the arrowheads appear to be different lengths when they are really the same length
- Mueller Lyer
Lines look like they are curved outwardly when they are actually parallel
Same as herring illusion but the lines appear to be curved inwardly
Parallel lines look slanted in opposite direction when you put diagonal lines across them
When we are analyzing our enviro we have to decide which parts of our enviro should be the figure (what to focus on) and which part should be the ground(background)
Vase vs faces; cow
Pictures of duck vs rabbit, young girl vs old women are examples of what illusion
reversible and multi-stable figures
Examples of theories that people came up with to explain illusions
Feature analysis (bottom up) - structuralism, visual alphabet/geons
Global processing (top down) - Gestalt organizing principles
Computational models - David Marr
To identify the smallest set of "primary sensations" necessary to uniquely ID any object. Ex: color, size, shape, texture (19th century)
Problems with structuralism
- couldn't agree on which senations were primary
- explosion of number of primary sensations that were identified (>40,000)
The perceptual ID of an object by the simpler components of compromising it. The preference of cortical cells for orientation is used to support feature analysis theories
This model features demons
Oliver Selfridge's pandemonium model (letter ID)
Problem with Feature theories
they don't explain how features of a pattern are organized. Context is important ex: A in cat
Uses the assumption that you can recognize 3D objects involving the use of 24 differnt 3D shapes or GEONs
Recognition by components (Beiderman, 1995)
Problem with recognition by components
people are poor at identifying objects in unusual angles or orientations
Edges(boundaries) that are percieved despite the almost lach of physical differences between chape and background are?
Name of a figure made by illusory contours.
What responds to a specific sphape and that shape created by illusory contours, but with longer response latency?
Neurons in IT-INFEROTEMPORAL cortex
Information processed at the lower levels without influence from higher levels.
Bottom up processing
Information at the higher levels *via guessing, predicting) may influence processing at lower levels
Top down processing
3 basic concepts of bottom up processing:
Top down process is driven by...
- visual attention
- the need to accomplish a certain task
We percieve objects as well organized patterns rather than separate components; Conveys a concept of wholeness
Gestalt Priciple of visual perception
gestalt is german for form or shape
How do we get the organization to impose visual organization on stimuli?
past visual experience
10 Gestalt organizing principles
- Pragnanz (good figure)
- Good continuation
- common region
- Context or meaning
- common fate
Every stimulus is interpretted inits simplist form. Example?
Pragnanaz (good figure)
Similar things are grouped together, ex: color, size , orientation and aspects of form
Connected points resulting in straight or smooth curves belong together.
Items near each other are more likely to group together than items widely separated
Ex: rows of circles and squares. Proximity overrides similarity
If two features appear to be part of the same larger region, they will be grouped together.
If two items are connected, then they probably belong together
Things from groups appear meaningful or familiar.
context or meaning
13 (inbetween letters you would call it a B, inbw #s thirteen)
We tend to complete figures even when part of the info is missing
Elements changing at the same time are seen as belonging together
Elements of a scene that share the same direction and speed of motion tend to be perceived as a unit
ex. dalmation and moving dots
Your ability to read a paragraph with letters out of order
Gestalt processing in reading and word recognition
Which popout is faster, top down or bottom up?
(in top down you have to search)
Different visual features are coded in parallel in separate feature maps.
Visual search is easy when it involves only a single feature that can be computed by a feature map
Feature integration theory (treisman & gelade, 1980)
Attention is needed at a particular location to synthesize its features into an object (serial search)
look for red circle in photo of red squares and blue circles
Assumes that items are examined one at a time, terminates when target is found/all items examined
If there is a target present how many items need to be examined in a serial search?
- ~ half
- n is size of search set
If there is a target absent how many items need to be examined in a serial search?
all items will be examined
How can different visual characteristics of an object be held together after different visual info is processed within diff neural channels, depending upong different pathways with different transmission and processing speeds?
AKA how do we put everything back together after we've taken it apart to process it?!