Lecture 6: Contours to Objects

Question 1

how do we get from contours to shapes?

Answer 1

V4 neruons have complex tuning that can detect changes in curvatures (convexes and concave). May have diff neruons that respond to different curvatures.

Question 2

what did they find whilst looking at population neurons of monkeys?

Answer 2

They looked at the response of neurons of monkeys towards different shapes and attempted to recreate the shape the monkey saw through looking at the population codes.

• Population of V4 neurons can represent complex shapes via their curvature preferences
• A more complex version of population code than we have seen in V1

Question 3

Illusory contours

Answer 3

we can see shapes even when there is not an explicit shape there which is a consequence of our visual processing assumptions

Question 4

top down visual processing

Answer 4

when looking at highly contrasted image does not provide information, but if we see the og image, we can then see shapes in the contrasted image - makes it more meaningful (integration of top down and bottom up processing)

Question 5

how do we process faces

Answer 5

we have specific neurons that specifically respond to peoples faces

Question 6

why are we so efficient at face perception?

Answer 6

• domain specificity: evolutionary mechanisms to process faces differently to normal objects
  expertise: we do it every day and become finely tuned to faces

Question 7

what way do we process faces?

Answer 7

holistically - representing features and their relationship as one unit (see eyes/mouth/nose etc as one unit).

Question 8

part whole effect:

Answer 8

Features are easier to identify when presented as part of a face

Question 9

Face inversion effect

Answer 9

Inversion of any kind of stimulus disrupts processing of fine details and relationship between features

Question 10

prosopagnosia

Answer 10

when there’s the difficulty of perceiving stimuli very specific to faces - failure to identify or distinguish between faces.

Question 11

norm based code

Answer 11

• Facial features (expression/properties etc) are represented as deviations from the average face (can be seen in image to right – norm to the middle and then faces around that is deviations from the norm).

Question 12

identification of caricatures

Answer 12

Caricatures have exaggerated what makes that identity that identity making it easier for us to recognise these caricatures – features become more distinctive.