WEEK 1 - Introduction and object recognition Flashcards
Ambiguous/reversible figures
in the brain = perception changes but not the stimulus
perception reverses but not visual information
perceptual constancy
the tendency of seeing an object still In its form despite changes in orientation
- shape constancy
- size constancy
- brightness constancy
Tachistoscope procedure
This device flashes a series of images on to a screen at a rapid speed to test visual perception, memory, and learning
Word frequency effect
we SEE more often, we recognise them more quickly
Recency effect
Going to recognise a word faster if we have just been exposed to it as opposed to not having seen it recently
Repetition priming
change in responding to a word or an object as a result of a previous encounter with that same item, either in the same task or in a different task.
-implicit memory
Word superiority effect
e.g more likely to recognise ‘boat’ rather than ‘b’ when presented at the same time)
Degrees of wellformedness
(e.g ‘Fike and and ‘Lafe’ as opposed to ‘xshs’ and ‘wynp’)
Feature net model
A system for recognizing patterns that involves a network of detectors.
Letters are identified via their component features”.
prosopagnosia
inability to recognise faces
- might occur due to brain damage or a condition since birth
- still able to identify shapes
Inversion effect
Faces are much harder to recognise when they are upside down, but does not apply on objects
holistic recognition
faces are processed as a whole and not purely based one its features
composite effect
Occurs when two faces are split horizontally and stuck together, our brain struggles to identify two operate identities as we recognise it as a whole
Introspection (Wundt & Titchener)
The examination or observation of one’s own mental and emotional processes
- Majority of mental activity occurs in our unconscious, even though introspection focusses on the conscious
- How people experience events is subjective to the individual
The transcendental method (Kant)
idealism. triangle thing
. (1) Ego transcendence
(self: beyond ego), (2) self-transcendence (beyond the self: the other), and (3) spiritual transcendence (beyond space and time).
Behaviourism (Watson)
focused not on the internal emotional and psychological conditions of people, but rather on their external and outward behaviors. He believed that a person’s physical responses provided the only insight into internal actions
Computers
can be characterized as the study of the behavior of computer systems, the kinds of personalities they exhibit, and the human reaction to these personalities and behaviors. Today there are numerous computer systems that have definite personality characteristics.
Gestalt laws of perceptual organisation
The human brain is exceptionally good at filling in the blanks in an image and creating a whole that is greater than the sum of its parts. It’s why we see faces in things like tree leaves or sidewalk cracks.
Word superiority effect
top down processing: words provide contact for letters so we dont need to actually read each letter
What is the feature net model of word recognition
A model for word recognition based on dectectors activating on three levels of perception: feature, letter & word
excitatory connection
link between nodes so that one node when activated then activates the other
recognition by components model
is a top-down process proposed by Irving Biederman in 1987.
- RBC theory
- we are able to recognize objects by separating them into geons
What is one advantage of object recognition based on geons?
Geons are viewpoint-independent: no matter what angle you are at, you are able to recognize the object
What are two differing theories of object recognition?
RBC: Recognition by components model (geons, viewpoint-independent) or viewpoint-dependent models
What are two top down factors influencing word recognition?
1) frequency (familiarity)
2) recency
What is the “degree of well formedness” principle of word recognition?
the more english like a string of letters, the easier it is to recognize it
What is the activation level of detectors in the feature net model based on?
1) recency 2) frequency
What is the importance of bigrams in the feature net model?
Bigrams - strings of two letters - provide evidence for the word superiority effect and help to reduce confusion due to limited or obscured input
What are the three (or four) levels of detectors in the feature net model?
feature, letter & word (plus proposed bigram in between letter & word)
What did the McClelland & Rumelhart Model add to the feature net model?
They proposed that detectors can both inhibit & excite detectors at higher levels and other detectors at the same level
What did the McClelland & Rumelhart Model add to the feature net model?
They proposed that detectors can both inhibit & excite detectors at higher levels and other detectors at the same level
What is an exception to both the feature net and the RBC (Recognition by Components) models?
Face recognition - its holistic, not component based
What is the difference between template matching and prototype matching in object recognition?
template matching is the belief (widely disproved) that we recognize objects by matching them to an image of what the object/letter should look like. Prototype matching on the other hand says we match letters to a prototype: an average of all category members (images) encountered
What is an example of why prototype matching is more generally agreed to be true?
Letters come in all forms, shapes & sizes and we still recognize them as the same thing
word superiority effect
refers to the phenomena os how people better recognise letters than words
Inversion effect
difficulty recognising upside down pictures
