3. Object Recognition Flashcards
Biederman’s recognition-by-components theory proposes that objects consist of primitives known as what?
geons
What series of stages are involved in object recognition?
- percieve/ see object
- structural description of the object
- semantic representation
- name representation
- name
Viewpoint invariant mechanisms of object recognition are usually involved when a task involves what?
a
Which disorder involves problems with object recognition that result from deficits in perceptual processing?
Apperceptive Agnosia
In which disorder can patients copy objects, even though they don’t recognise them?
associative agnosia
Differences between face and object recognition?
- Object recognition tend to involve between category discriminations
- face recognition always to involve with-category discriminations
What are the challenges with intra-class variation for object recognition?
- We can still recognise something (e.g. chair) even though the features and appearance vary vastly.
- Recognition when only part of an object is visible.
What are the challenges with viewpoint variation for object recognition?
- Is more difficult to make discrimination if it is the same person across different view points
- Even though we might not be able to tell if it is the same person we can still tell that it is a person.
What are the 4 theories of 2D pattern recognition?
- Template theories
- Prototype theories
- Feature theories
- Structural descriptions
Describe the Template theory of 2D pattern Recognition.
- We have a mini copy or template in LTM of all known patterns.
- Compare stimuli to templates in memory for one with greatest overlap until a match is found
- Recognise an object when there is the most overlap from templates form LTM
What are the problems with the Template theory of 2D pattern Recognition?
- Problem of imperfect matches
○ Cannot account for the flexibility of pattern recognition system - Cannot account for our flexibility in ability to recognize items.
- Comparison requires identical orientation, size, position of template to stimuli
Describe the Prototype theory of 2D pattern Recognition.
- Modification of template matching (flexible templates) = one average template stored in memory.
- Possesses the average of each individual characteristic
- No match is perfect; a criterion for matching is needed
What is the evidence for the Prototype theory of 2D pattern Recognition?
Franks & Bransford (1971) ○ Presented objects based on prototypes ○ Prototype not shown ○ Yet participants are confident they had seen prototype ○ Suggests existence of prototypes
What are the problems with the Prototype theory of 2D pattern Recognition?
○ Old literature so not too reliable (i.e. Franks & Bransford, 1971)
○ Consistent with prototype theory but doesn’t discount other possibilities
Describe the Feature theory of 2D pattern Recognition.
- Pattern consists of a set of features or attributes.
- A = 2 straight lines & connecting cross bar.
- But also need to know relationship between features as well as the features themselves.
Describe the Structural descriptions theory of 2D pattern Recognition.
“describes the nature of the components of a configuration and the structural arrangement of these parts” (Bruce & Green, 1990)
What are the problems with the Structural descriptions theory of 2D pattern Recognition?
- Need to know relationship between these features as well
Describe the process of 3D object recognition.
- interpret input to visual system as coherent structures, segregated from one another & from background. (Early image processing)
- create structural description of the object
- which can then be matched to the descriptions of visual objects stored in memory.
What did Marr & Nishihara (1978) propose?
○ Can break all objects down into a series object cylinders (feature/ components)
○ Need to work out the relationship between cylinders. This creates your structural description of that object.
○ expressed structural relations by a hierarchical organisation of cylinders.
What did Biederman (1987; 1989) propose?
- Recognition-by-components theory: Objects composed of basic shapes
○ GEONS = ‘geometrical ions’.
○ approximately 36 different volumetric shapes
○ Viewpoint invariant theory
How can geons vary?
- Relative size
- Verticality
- Centring
- Relative size of surfaces at join
What are the ‘non-accidental’ properties of geons?
○ Curvature - points on a curve
○ Parallel - set of points in parallel
○ Co-termination - edges terminating in a common point
○ Symmetry - versus asymmetry
○ Co-linearity - points in a straight line
What is the object recognition process that we go through according to Biederman (1987)?
- Edge extraction
- Detection of Non-accidental properties
or - parsing of regions of concavity
- Determination of components
- matching of components to object representations
How did Biederman (1987, 1989) evidence his theory?
- deleted edges of line drawings at points where easily reinstated or difficult to determine.
- Deletion of component affects matching stage
- Midsegment deletion makes it more difficult to determine components.
- Midsegment deletion led to less errors than component deletion.
What is the support for Biederman’s geon theory?
- Vogels, Biederman, Bar & Lorincz (2001) found some cortical neurons in monkeys sensitive (responded more) to geons.
What are the disadvantages of Biederman’s theory?
- Structure is not always key to recognition: Peach vs. Nectarine
○ Colour is important here not just geons
○ Face discrimination: texture, colour are key for identification - Simplifies the contribution of viewpoint-dependence
Describe the Viewpoint Dependent Theory.
- Assume changes in viewpoint reduce the speed and / or accuracy of object recognition.
○ Suggest viewpoint matters. - Object representations are collections of views that depict the appearance of objects from specific viewpoints’
- Viewpoint dependent more important for within category discriminations
What did Vanrie et al. (2002) propose about Viewpoint dependent and invariant theories?
- Viewpoint dependent = complex within category decisions
- Viewpoint invariant = easy categorical decisions
Both views could be correct just in different contexts
what is the binding problem?
how do we integrate different kinds of information to produce object recognition?
What is the process of object recognition?
- object
- structural description
- semantic representation
- name representation
- name
- all or nothing model - you cant get to the next stage if you get stuck at an earlier stage
Describe Humphrey’s cascade model?
- structural, semantic and name stages interact.
- More of interactive account - not an all or nothing account
- Get inhibition and interaction between stages
- problems in earlier stage have knock-on effects with later stages, they don’t halt the process all together.
what is Agnosia?
- Failure of knowledge or recognition
- In visual agnosias, feature processing and memory remain intact, and recognition deficits are limited to the visual modality.
- Agnosia restricted to specific categories
Describe Apperceptive Agnosia.
- Problems with early processing (shape extraction).
○ All objects may look the same. - Perceptual deficit, affects visual representations directly, components of visual percept are picked up, but can’t be integrated, effects may be graded, often affected: unusual views of objects.
What is the test for Apperceptive Agnosia?
- ability to recognize degraded or altered-view stimuli is impaired.
- perceptual deficit - problem in early processing.
What is Associative Agnosia?
- Problems with later processing (recognition).
- Visual representations are intact, but cannot be accessed or used in recognition. Lack of information about the percept.
- Can copy drawings of objects but cannot name them
