L11 - Object Recognition Flashcards

1
Q

What is object recognition?

A
  • The process of matching a representation of a viewed object to a representation stored in memory.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is recognition by Template Matching?

A
  • Perceived stimuli are compared to templates stored in memory.
  • If they match the object is recognised.

Problematic however, when there is variation in stimulus types, changes in viewpoint, illumination.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are the 3 stages of object representations conceptualised by David Marr?

A
  1. Primal sketch:
    - primitive but rich description of the way light information changes across the visual field.
    - captures visual edges and vertices.
  2. 2½-D Sketch:
    - recovery of surface information.
    - from the point of the observer (viewer-centred).
    - doesn’t generalise to other viewpoints.
  3. 3-D Model representation:
    - recovery of general volumetric parts that make up the object.
    - independent of viewing point (object-centred).
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is Irving Biederman’s Recognition by Components Theory?

A
  • All objects are made up of volumetric components (geons).
  • 36 geons are sufficient to describe all objects.
  • Objects are represented by specific geons and the relations between them.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the series of steps in Bierderman’s Theory?

A

Edge extraction (similar to Marr’s primal sketch) -> Parsing regions of concavity OR Detection of non-accidental properties -> Determination of Components (geon decomposition) -> Matching components to object representations.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is the View-Based Theory of Object Recognition?

A
  • Objects are recognised based on stored views that we have encountered through our experience.
  • New views must be matched to stored views.
  • This requires some form of transformation into alignment.
  • “Views” are holistic representations.
  • Have difficulty accounting for our ability to categorise new objects as members of a category.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly