lecture 4 Flashcards
(29 cards)
describe template matching
• Example: letter recognition
• Whole stimulus / pattern of excitation on retina is matched against a template in memory (template memory representation of typical instance of an object)
• What to do with slight deviations in shape, size, orientation?
– Normalization of input pattern
describe feature comparison
Analyze retinal image for presence/absence of
certain features
• Features give evidence for a certain object
(e.g. a certain letter)
what are navon stimuli
– Global letters made up of local features
– Global precedence effect
• Decisions about small letters were slower if large letter was different
• Decisions about large letters were not slower if small letters were different
What is the limit of template matching
too dependent on normalization
describe perceptual segregation
– Separating visual input into individual objects
– Thought to occur before object recognition
describe the gestalt psychology
The Law of Prägnanz
• “Of several geometrically possible organisations
that one will actually occur which possesses the
best, simplest and most stable shape”
(Koffka, 1935, p. 138)
what are the gestalt laws of perceptual organisation
a) The law of proximity
b) The law of similarity
c) The law of good continuation
d) The law of closure
describe uniform connectedness
Palmer and Rock (1994) proposed the principle of
uniform connectedness:
– Any connected region having uniform visual properties (e.g., colour, texture, lightness) tends to be organised as a single perceptual unit
– Occurs before and can overpower Gestalt grouping
laws such as proximity and similarity
describe why figure–ground segregation is innate fails
But amnesic patients do not show awareness of shapes of familiar objects in silhouette in a task where they identified which was the figure in figure-ground combinations
describe biedermans recognition-by-components theory
Objects consist of combinations of geons
– Geons = geometrical ions
– +/-36 basic shapes
object is for example a telephone or suitcase etc.
describe edge extraction
Five properties of edges that are invariant across viewing angles: • Curvature – Points on a curve • Parallel – Sets of points in parallel • Cotermination – Edges terminating at a common point • Symmetry – Contrast with asymmetry • Collinearity – Points sharing a common line
describe a limitations Recognition-by-components
• De-emphasises importance of contextual influences,
expectations and previous knowledge
• Fails to account for most within-category discriminations
• Much recognition is actually viewpoint-dependent
• Some classes do not have invariant geons yet are still
recognisable as members of a category (e.g., clouds)
what is the influence of viewpoint
- Categorisation of objects (between category; e.g. is the object a dog) does not depend on viewpoint
• Identification of objects (within category; e.g. is the
object a poodle?) does depend on viewpoint
describe the face inversion effect(holistic face processing)
Inverted faces are disproportionately harder to recognise than upright faces relative to other objects
describe the part whole effect(holistic face processing)
Memory for face parts (e.g. mouth) more accurate when presented within the whole face; little difference for house parts (e.g. door)
describe the composite effect(holistic face processing)
Deciding whether the top half of a face was the same or different as in a previous composite face is more difficult when it appears above a different bottom half than above the same bottom half
describe why face processing is impaired in prosopagnosia
lateral fusiform gyrus responds morte to upright face than inverted faces or object and is often damaged in prosopagonia
Describe bruce and youngs model –> structural encoding
Various representations or descriptions of faces
Describe bruce and youngs model –> expression analysis
An emotional state can be inferred from facial
features
Describe bruce and youngs model –> Facial speech analysis:
Speech perception aided by lip movements
Describe bruce and youngs model –> directed visual processing
Specific facial information may be processed
selectively
Describe bruce and youngs model –> face recognition units
Structural information about known faces
Describe bruce and youngs model –> person identity nodes
Information about individuals (e.g., occupation,
interests)
Describe bruce and youngs model –> name generation
A person’s name
