Possible processes in pattern recognition? (3)
- Template match (implausible)
- Constituent features
- Pattern identified by associated features* (may/may not be present every time; better than constituent (would still recognize an A w/ no peak)).
3 Steps in object recognition from visual features
- Features detected in parallel
- Feature combos matched to stored info
- Object identification based on best match
Recognition Pandemonium (better than object rec. from visual features)
- Recognition of alphanumeric characteristics
- Features are line segments and intersections
- Independent detectors for each pattern
- Most activated pattern is recognized
Problem with Feature Theory
No general account of features and how they are detected or combined.
Synaesthesia
An involuntary, consistent phenomena in which stimulus are experienced in a different perceptual pathway.
A type of Synaethesia where colours are elicited by letters or digits, though the association differs between subjects
Graheme-Colour Synaesthesia
Stroop Task
- name the ink colour of colour words, or read the colour word
- Naming inconsistent colours is difficult but the reverse is less problematic
Interpretation of the Stroop Effect
- Words automatically read
- Word reading interferes or competes with colour naming
- Reverse Stroop effect is weaker
How can we test Synaethesia with the Stroop Task?
- Colour digits consistently or inconsistently w/ synaesthetic association
- Asked to name colour of digit
- synaesthetic person does worse on the incongruent task, but there is no difference for a normal person
Evidence that Synaethetic “cross connection” isn’t purely perceptual (and is also conceptual)?
-when presented w/ a digit, operation, digit, colour patch, then asked to name colour, and then asked to report the arithmetic solution,
the conceptual colour associated with the digits interferes w/ colour naming in synaesthetics.
