Visual Perception

Question 1

Constructivist approach

Answer 1

Top-down (extraction of information): Herman von Helmholtz: History.

Interaction between internal and external sources of information - expectations can lead to errors (visual illusions).

Retinal size does not lead to actual size, but picture in relations to other things does (inferred from cues).

Knowledge is needed to make two dimensional images three dimensional.

All the size cues are from a constructivist approach

Question 2

Convergence (binocular cues)

Answer 2

The eyes converge (move together) to look at nearby objects.

Binocular disparity: distance of eyes guarantee that you will see two slightly different things when objects are in different distances - stereopsis

Question 3

Accommodation

Answer 3

Lens changes according to distance of object:

thin - distant

thick - nearby

Question 4

Interposition

Answer 4

If objects cover one another, we judge the partly covered object to be further away.

Question 5

Size cues

Answer 5

Object size relative to each other tell us about distance (ie cow close vs cow far away)

Question 6

Linear perspective

Answer 6

parallel lines seem to meet in distance

Question 7

Atmospheric perspective

Answer 7

distant objects look blurred and bluish

Question 8

Shading

Answer 8

Cue provided from shadow and light, tells you which object is closer (can produce misleading images)

Question 9

Height cues

Answer 9

objects close to horizon appears further away

Question 10

Kinetic depth effect

Answer 10

a figure that looks static appears to have depth when it moves

Question 11

Evidence for constructivist approach

Answer 11

ie Eistein’s face, Muller-Lyer (‘angles in’effect) - optical illusions

Cultural differences in expectancies (ie only people with experiences with corners will see Muller-Lyer effect) - experience is a factor.

Question 12

Ecological approach

Answer 12

James Gibson. Challenges constructivist approach:

Main purpose of vision is to facilitate interactions between observer and environment - Bottom-up processing

If environment is perceptually rich and we can move in it, there is no need for processes of interference in order to extract information.

Objects = affordances (ie a ladder affords up and down

Access to information through resonance

Transformational invariants = aspects of perception that persists over time and are left unchanged despite changes in retinal image.

Question 13

Texture gradient

Answer 13

Texture of surface becomes denser with distance.

ie Visual cliff

(Ecological)

Question 14

Motion parallex (Optic flow)

Answer 14

Objects at different distances appear to move in different directions and speed.

(Ecological)

Question 15

Dorsal stream (perception)

Answer 15

Perception for action: how do I interact with what I see

Question 16

Ventral stream (perception)

Answer 16

perception for identification: What do I see?

Question 17

reconciliation of constructivist and ecological approach

Answer 17

Norman: dorsal often works without conscious awareness (=underpins ecological approach), ventral higher system that works based on conscious outputs (=underpins constructivist approach)

Question 18

Model of visual object recognition (ventral)

Answer 18

Riddoch and Humphrey model: grouping elements into higher order units:

ie object–>form features etc

Question 19

Gestalt principles of grouping

Answer 19

principles to explain the groupings

Question 20

The law of proximity

Answer 20

visual elements more likely to be grouped if close together

Question 21

The law of similarity

Answer 21

visual elements grouped together if they share visual attributes (ie shape/colour)

Question 22

Law of good continuation

Answer 22

edges are grouped together to avoid changes/interruption

Question 23

Law of closure

Answer 23

missing parts are filled in (important for recognising partially occluded objects) (think the triangle in shaped by ‘packmans’)

Question 24

Law of common faith

Answer 24

elements that move together tend to be grouped together (think dots in the dark –> man walking)

Question 25

Neuropsychological evidence

Answer 25

integrative agnosia patients (difficulties integrating parts into wholes - Can describe a carrot, but doesn’t know what it is and can’t see hidden objects as wholes etc)

Question 26

Structural descriptions: Object constancy

Answer 26

= recognising object across viewpoints and light conditions - based on stored information in memory

Question 27

How does the brain obtain object constancy?

Answer 27

Marr & Nishihara: objects decomposed into segments (cones), axes extracted - this can be done regardless of viewing point

Foreshortening views (ie a car seen from the front) more difficult to determine axis (evidence as this is what some agnosic patients struggle with)

or

Biederman: objects consist of basic shapes (geons) which are extracted from objects regardless of position (evidence: object orientation agnosia where patients are unable to decide whether two objects have the same orientation)

Both these probably play a part in perception

Question 28

visual cues for coding of object properties (dorsal)

Answer 28

binocular vision & motion parallax especially important

Question 29

Understanding object’s function

Answer 29

Sometimes understanding is based on shape etc (patients who didn’t know what scissors were still held it by handle) –> ie sensory-motor transformation based on affordance

Question 30

Imitation and action understanding

Answer 30

sensory-motor mapping system (based on other humans) mirrors visual inputs and transforms it to executions - mirror neurons!