object and face recognition

Question 1

figure ground relationship?

Answer 1

Perceptual grouping which is essential in recognising objects
Top down

Question 2

Law of Pragnanz

Answer 2

• Notion that we aim to see SIMPLEST possible organisation of visual environment
• why we don’t often notice little mistakes

Question 3

Gestalt laws that come under the law of pragnanz - ie making it possible for us the see the simplest form (4)

Answer 3

1) Law of PROXIMITY: close things will be grouped
2) Law of SIMILARITY: similar things will be grouped
3) Law of good continuation: groups things which create LESS CHANGE in flow
4) Law of CLOSURE: Missing parts of an object will be filled in

Question 4

Marr’s computational theory of perception and vision

Answer 4

PRIMAL SKETCH:
a) Raw sketch: place tokens ie bars and edges
b) Fill in these edges: group them
~ many ambiguities: explicit naming and least commitment principle is used ~
2.5D SKETCH:
a)depth, texture, law of common fate
3D sketch:
a) Arranged hierarchically
b) matched with real world knowledge
c) shapes recognised due to their CONE’S AXIS

Question 5

Explicit naming

Answer 5

giving name to a set of grouped elements/ symbols

this name can be used over and over to describe other sets of grouped elements

Question 6

Least commitment principle

Answer 6

says ambiguities are resolved only when convincing evidence as to the appropriate solution

Question 7

Biderman (1987)

What & how do they work?

Answer 7

GEONS: extension to Marr
What: 2d/3d simple forms of shapes (cylinders/ cubes/wedges etc)
How does it work: during recognition they’re extracted & matched against structural descriptions we have STORED MENTALLY = if good match found = recognition

Question 8

How are faces remembered?: Featural hypothesis

Answer 8

faces are remembered due to their features

Question 9

How are faces remembered? configurational hypothesis

Answer 9

Faces are remembered due to the relationship between facial features

Question 10

face inversion effect: what and why?

Answer 10

• faces are harder to identify when upside down
• why?:
  a) EXPERTISE OR
  b) combination of featural & config hypothesis (when inverted holistic processing cannot occur)

Question 11

Prospopagnosia

Answer 11

inability to recognise familiar faces

no FRUs/PINS

Question 12

Theoretical approaches to face recognition and processing: BRUCE & YOUNG (1986)

Answer 12

FUNCTIONAL FACE RECOGNITION MODEL

1) encoding of faces
2) face recognition unit activated (FRU): structural info about faces
3) if theres a match betwwen encoding and FRU -> Semantic info accessed
4) Personal identity node (PIN) accessed - personal info about person

Question 13

Theoretical approaches to face recognition and processing: (Burton, bruce & young: 1993)

Answer 13

Added stuff to the model:

1) Name recognition units
2) Word recognition units: associated words
3) semantic info units: personality/occupation

Question 14

Capagras Syndrome

Answer 14

-can identify at name level but see an an imposter

low activation of FRUs

Question 15

Emotion & face recognition

Answer 15

we identify HAPPIER faces faster than NEGATIVE ones

Question 16

Greebles

Answer 16

Gauthier & Tarr (1997):
Can train someone to be greeble expert in matter of hours
Use config hypothesis, relationship of their features
~ shows an innate preference but pref is REFINED