objects and scenes pt 1

Question 1

what is the inverse projection problem

Answer 1

2d image on retina from 3D world

• different objects can look identical (infinity -> 1)
• single object can look different (1->infinity)
  ex. ambiguous cylinder illusion

Question 2

viewpoint invariance

Answer 2

ability to recognize an object regardless of viewpoint

ex. same person from diff angles

Question 3

gestalt approach of principals of perceptual organization

Answer 3

• the whole is other than the sum of its parts
  ex. apparent motion, subjective color, illusory contours
  triangle and 3 circle illusion

Question 4

principles of perceptual grouping

Answer 4

good continuation: lines follow the smoothest path
good figure: patterns are as simple as possible
similarity: similar things grouped together
proximity: nearby objects appear grouped together
common fate: elements moving together grouped
common region: same region of space
uniform connectedness: connected regions single unit

Question 5

principles of perceptual segregation

Answer 5

figure ground seperation:
- figure is thing and memorable
- ground is more uniform and not memorable
border belongs to the figure (border ownership)

Question 6

figure features

Answer 6

lower and convex

Question 7

gestault approach features

Answer 7

holistic- way we interpret objects as the whole

bottom up- relying on the stimulus and not past experience

Question 8

RBC theory

Answer 8

recognitino by components theory

• we percevie elementry features to observe objects
• an object is recognized when enough info is available to identify its geons

Question 9

geon

Answer 9

set of canonical shapes that we can think of the world as being composed of

• discriminable
• resitance to visual noice
• invariance
• distinct

Question 10

principal of componential recovery

Answer 10

object recognition is not based on the amount of information but the ability to identify its geons
- can see object best when u have the most info about geons (corner/turns)

Question 11

JIM (jim and Irvs model)

Answer 11

builds up from small details to an overall object

• bottom up, neurally inspired
• edge detectors/vertices

Question 12

what is the GIST of a scene

Answer 12

counters the RBC theory which says we build up from small details to overall objects
experiments shown that we perceive large scale properties first and then smaller

Question 13

how do we perceive gist

Answer 13

global image features- naturalness, openness, roughness, expansion, color

• simultaneously process visual scene at frequencies
  low: gist
  high: detail

Question 14

hybrid image

Answer 14

low spatial frequencies: mairlyn monroe
high spatial frequeuncies: albert einstein
- conflicting info at high and low spatial frequency changes perception

Question 15

top down processing

Answer 15

perception is not purely based on stimulus

depends on experience, expectation and goals

Question 16

experience

Answer 16

figure ground seperation- meaningfulness influences seperation
perceptual organization- based on scenic patterns
object recognition- identify objects based on context

Question 17

hemholtz theory of unconscious interference

Answer 17

likelihood principle - perceive the object most likely to cause pattern of stimuli
unconscious inference- application of likelihood principle is unconscious but based on past experience

Question 18

bayesian inference

Answer 18

interpreting situations in a way that is most likely based on what you know and what you currently are perceiving
-combines posterior belief/ prior belief/ likelihood