Transcript Pattern Recognition

Pattern Recognition
Pattern - complex composition of sensory
stimuli that the human observer may
recognize as being a member of a class of
objects
Issue - what cognitive mechanisms need to be
inferred to describe this process of
recognition?
Bridge with Signal Detection
• Detection of sensory stimuli - data driven
• Perception of Patterns - conceptually driven
• work from the bottom (identifying stuff in
the world) to the top (thinking)
Necessary Terms and Concepts of
Pattern Recognition
• Serial and Parallel Processing
– Serial or sequential processing means we
process information one step at a time, where
one process must be finished before the next
can be started.
– Parallel processing means we can process
several tasks at one time
Necessary Terms and Concepts of
Pattern Recognition
• Bottom-up and Top-down processing
– Bottom-up processing is similar to inductive
reasoning. Basic data are combined into more
complex forms.
– Top-down processing is similar to deductive
reasoning. Higher levels of processing affect
lower level tasks.
• The following gives examples of how we perceive
visual patterns and how positioning or additional
information affects our perception.
Theories of Perception
1. Gestalt
(Canonic Processing)
2. Bottom-Up vs. Top Down
3. Template matching
4. Feature Analysis
Prototype Theory
Form Perception
Gestalt Theory
• Gestalt theorists are among the earliest to
look at the problem of pattern recognition.
• They postulate that we perceive stimuli as a
whole pattern. That is, that individual parts
have no meaning independent of the whole
but combine to revel an identifiable pattern.
• Gestalt theorists developed 5 rules of
perception to explain their ideas...
Gestalt Laws
1. Law of Proximity:
– Elements that are
closer together will be
perceived as a coherent
object.
– On the top, there
appears to be three
horizontal rows, while
on the bottom, the
grouping appears to be
columns
Gestalt Laws
• Law of Similarity:
– Elements that look
similar will be
perceived as part of the
same form.
– There seems to be a
triangle in the square.
Gestalt Laws
• Law of Closure:
– Humans tend to
enclose a space by
completing a contour
and ignoring gaps in
the figure
Gestalt Laws
• Law of Prananz:
– A stimulus will be
organized into as good a
figure as possible
(symmetrical, simple,
and regular)
– The figure appears to the
eye as a square
overlapping triangle, not
a combination of several
complicated shapes.
Summary of Gestalt
• Modern conclusion is that some level of
“natural organization” of patterns is tied to
the perceptual history of the subject
– a function of the perceiver rather than the
stimulus
Canonic Processing
• Extension of Gestalt
• the first images of an object that comes to
mind when thinking of that particular form.
• perspectives fluctuate with culture and time.
– person from Los Angeles asked to think of a
house might recall a one story, 3 bedroom
stucco structure; a person living in a povertystricken Third world country might imagine a
small hut made of tree branches held together
with mud
Canonic Processing
• Through common experience with objects,
we develop memories of the most
representational view (and gives most
amount of info)
• Studying this helps to understand form
perception, prototype formation, economy
of thinking, “visual shorthand”
Top-down vs. bottom up
processing
• Bottom-up processing consists of mental
operations influenced by the physical
properties of the stimulus.
• Top-down processing consists of mental
operations influenced by the results of
processes already completed.
• Reading the following requires both kinds
of processing:
Reminder...
• Problem - perception requires that
information in the environment must be
matched to internal information about the
environment; however, the environmental
information is subject to substantial
variation. How do we recognize things in
the face of this variability?
Template Matching
• Template - internal constrict that, when
matched by sensory stimuli, leads to the
recognition of an object
• Assumption: a retinal image of an object is
faithfully transmitted to the brain and that
an attempt is made to compare it directly to
various stored patterns
Template Matching
• compare stimulus to large number of literal
copies (templates) that are stored in
memory to find match against all templates
– works well with computers (check-sorting
machines)
– does not work well with humans -- too
inflexible
• does not account for similarities among objects
• what is the effect of context?
Prototype Model
• more flexible version of template model the match does not have to be exact
– match against “prototypical A”
• advantages
– manageable number of representations in
memory can account for how people classify
similar objects into a common category
• disadvantages
– lack of explicit information about how stimuli
are compared to prototypes
Feature Analysis Model
• Assumption: stimuli consist of
combinations of elementary features; (e.g
for the alphabet, features may include
horizontal lines, vertical lines, diagonals,
and curves)
– make discriminations based on a small number
of characteristics of stimuli
– distinctive feature components stored in
memory [a mini-template model??]
Feature Analysis Model
• What is a feature?
– A feature is a distinctive attribute or
characteristic of a stimulus.
e.g., 'T' has 2 features: ' - ' & '|'
(E. Gibson, 1969)
Feature Analysis Model
• Psychological Evidence: Gibson (1969)
• decide whether or not two letters are
different
• takes longer to respond to P & R versus G
&M
• P & R share many critical features
Feature Analysis Model
• Neurological Evidence: Hubel & Wiesel
(1962)
– microelectrodes in cats’ brains (visual cortex)
– some neurons respond only to horizontal lines,
others to diagonals...
– similar evidence in monkeys (Maunsell &
Newsome, 1987)
• certain feature detectors are “wired” and
help us identify features and simple patterns
Neisser example
• - Look for the “X”
OOPOPOPOP
POPPOPPPO
OOPPOXPOP
OOPOPOPOP
POPPOPPPO
Neisser example
• Look for the “X”
NNZNZNZNZ
ZNZZNZZNN
NNNZNXNZN
NNZNZNZNZ
ZNZZNZZNN
Feature Analysis
• advantages
– economical to store features in memory
– experimental evidence consistent with features
• disadvantages
– lack of applicability to a wide range of stimuli
– analysis of stimuli does not always begin with
features
– treats all features as equivalent
Back to Gestalt...
Back to Gestalt...
Back to Gestalt...