Object Recognition

Question 1

2 pathways of the brain

Answer 1

ventral and dorsal

Question 2

what is object agnosia

Answer 2

intact primary visual functions, verbal descriptions and logical reasoning but immediate object recognition is lost

Question 3

which stream is damaged in object agnosia

Answer 3

ventral

Question 4

What kind of approach is Marr’s model of object recognition/vision

Answer 4

computational

Question 5

What does Marr’s model of vision suggest

Answer 5

1. vision is an information processing task
2. we have to understand the nature of the task
3. we have to understand how it can be accomplished
4. we need to propose algorithms and mechanisms that can accomplish the visual task
5. explanations of visual experience and visual physiology should come from an understanding of the implementation of those algorithms

Question 6

Different models of object recognition used

Answer 6

template-matching model
feature-detector models
structural-description models
view-dependent models

Question 7

template-matching model

Answer 7

match the exact pattern of light on the retina with previously remembered patterns.
Neuron makes connections with pixels and when the specific pixels are detected the neuron fires

Question 8

problems with template-matching model

Answer 8

1. very constrained - can’t cope with diversity of natural object variation
2. doesn’t explain how we detect variances in images
3. there would be too many template detectors

Question 9

what model solves the problems of the template-matching model

Answer 9

feature-detector model: looks at specific features in the image rather than pixels

Question 10

Who put forward the idea of ‘feature-detectors’?

Answer 10

Hubel and Wiesel

Question 11

what do feature-detector models suggest?

Answer 11

we can detect specific combinations of features rather than specific patterns
the features don’t have to exactly match a pre-existing template, it just needs to contain the same sub-features that

Question 12

what are demons in the feature-detector model

Answer 12

subroutines

Question 13

who out forward the structural description model

Answer 13

Marr & Nishihara (1978)

Question 14

What did Marr & Nishihara (1978) suggest the goal of vision is in the structural description model

Answer 14

to describe the object unambiguously in its core geometric components

Question 15

Marr & Nishihara’s critera for good representation of high level vision

Answer 15

1. Accessibility
2. scope
3. uniqueness
4. stability
5. sensitivity

Question 16

what did Marr & Nishihara (1978) suggest the primitives (basic units of information in representation) are

Answer 16

objects are described in terms of axes and the volumes around them
the description is hierarchical
this means the object can be described at many scales

Question 17

according to Marr & Nishihara (1978) how is information organised into an abject description?

Answer 17

recognition:

by finding the closest model in the model store and specifying its parameters

Question 18

What did Biederman (1987) suggest for structural description models

Answer 18

similar to Marr & Nishihara (1978) but proposed a specific “alphabet” of primitive volumes callled Geons which object are decomposed

Question 19

what properties are geons similar to

Answer 19

any 2D projection

Question 20

how many geons are there estimated to be

Answer 20

less that 36

Question 21

evidence for structural models

Answer 21

intersections of geons are particularly informative in object recognition e.g. it is hard to tell what an object is when straight lines or corners are removed

Question 22

what are IT cells

Answer 22

higher visual neurons in IT area

Question 23

Pros of structural description models

Answer 23

explain invariance well
recognition is description not matching
evidence that structural information matters to both humans and neurons

Question 24

cones of structural description models

Answer 24

extracting model parameters can be hard in real mages
structural description is difficult for some objects e.g. campfires
driven by theoretical desirability rather than behavioural evidence

Question 25

what does the the view-dependent model suggest

Answer 25

that the arbitrary image has a brute association with the object

Question 26

what are the primitives in the view-dependent model?

Answer 26

sub-regions of the image

consisting of lines, curves, textures, colour, shading etc

Question 27

what kind of model is the view-based model

Answer 27

simple feet forward activation

Question 28

evidence for view-baed models

Answer 28

human object recognition is not perfectly viewpoint invariant
Monkey IT neurons selective for particular viewpoints
simulations show good invariance and similar errors in human observers

Question 29

pros of view-dependent model

Answer 29

straightforward
newer models based directly on what we know of physiology
abstract feature units are recombines
good evidence

Question 30

cons of view-based models

Answer 30

humans often show good generalisation across viewpoints even for novel objects
more memory intensive than other models e.g. geon models
no understanding of underlying relationships