Perception Flashcards
define
sensation
how our sensory receptors receive and represent stimuli in our environment
define
perception
how our brain organizes and interprets sensory information
what do visual illusions imply about perception?
it is an active process that our minds perform automatically
define
top-down processing
knowledge and expectations influence information-process
define
bottom-up processing
raw visual information is processed as is
provide an example of top-down processing
possible answers:
1. Bhalla & Proffitt (1999) found that wearing a heavy backpack influenced the perceived slope of a hill
2. wearing swim flippers makes swimming distances seem smaller
3. the color of food affects taste.
how do we see 3D objects when the image on our retinas is 2D?
- the retina contains cells that interpret basic information (eg edges, orientation of lines)
- the optic nerve extends information to the primary visual cortex (V1)
where is the primary visual cortex?
the back of the brain
fill in blanks with these words: left visual field, left visual cortex, right visual field, right visual cortex
information from _____ crosses and is processed in the _____________
- information from left visual field crosses and is processed in the right visual cortex
- information from right visual field crosses and is processed in the left visual cortex
what are occluded and granular objects? how does our brain register these objects?
depth cues
occluded objects are farther back. more granular (detailed) textures are closer. an object of fixed size is bigger when it is closer.
what are perceptual constances?
how our perceptions are stable despite changes in environment or context. i.e lightness, size, shape, color
how can our brain “see” things as normal when the image on the retina varies?
- assumptions about light sources (like direction, color, etc.)
- information from memory
- we don’t entirely know
the processes involved in building up a representation of an individual volume of matter in the world
1. indentifying and classifying edges
2. integrating edges
3. filling in missing edge and surface info.
4. unifying object parts
5. segmenting object from background
what did hubel & wiesel study?
hint: edge detection
they study single cells and found greater activity from cells that only responded to only one particular orientation of line
what is segmentation and grouping?
the brain differentiating which segments are part of the same object and which are diffferent
- must separte figures from ground
define
bistable perception
multiple interpretations of one image are possible
- an example of perception being an active process
define
proximity (Gestalt)
things that are closer to one another tend to be perceived as connected and part of a whole object, compared to things that are furhter apart
define
similarity (Gestalt)
things that are similar in brightness, color, shape, etc. tend to be seen as a single object, rather than separate objects
define
closure (Gestalt)
incomplete objects tend to be perceived as wholes
- the parts that form a complete/almost complete object go together
define
good continuation (Gestalt)
elements arranged on a line or a curve are perceived as more related than elements not on a line or curve
- i.e Olympics rings
define
figure-ground (Gestalt)
elements perceived as either figures (distinct element of focus) or ground (background where figures sit)
- i.e FedEx
Law of Pragnanz (Gestalt)
elements are grouped in the simplest way that produces a ‘good figure’
define
pattern recognition
ability to identify objects (matching patterns to knowledge)
how do we recognize objects?
identify the two steps object recognition
_TM and _TM
STM: represent stimulus in short-term memory
LTM: find representation in long-term memory that matches it
what are the three main object recognition theories?
- templates
- feature detection
- structural descriptions
define
template theory
a new stimulus is matchd to an internal mental representation of the stimulus (the template)
what are the issues with template matching theory?
viewpoint-dependent:
large variability within categories, requiring a new template for each possibility
define
feature detection theory
a stimulus is borken down into its features/components to determine what it is
describe how feature detection theory might explain how we can recognize an A or a dog
A: two long diagonal vertical lines, point at top, one shorter horizontal line
dog: two ears, four legs, a tail
what is the pandemonium model?
- instance of the feature detection theory
- Lindsay & Norman 1972
- suggests that feature ‘demons’ shout when they see their own feature
- ‘demon’s are more like populations of neurons
- primary visual cortex: feature demons
- cortical areas that determine object properites: cognitive demons
what is the word superiority effect?
people can more accurately identify letters when they are part of a word than on their own
- example of top-down-processing: knowledge of words influences perception of letters
what are some advantages of the feature detection theory?
- doesn’t require a large number of templates and invariant to size, orientation and location -> much more effficient
- explains errors people make during recognition (when images are presented too quickly, ‘B’ demons will shout as loudly as ‘R’ demons which may lead to letter confusuion)
- fits with neural evidence that shows neurons in the primary visual cortex act as feature detectors
what are the shortcomings of feature detection theory?
- difficult to define a “feature” (is the feature the letter A or the lines, points, etc)
- can’t explain the word superiority effect (increase in efficiency when letters are in a word)
- entirely bottom-up
define
recognition by components theory
stimuli are composed of universal features called geons (36 total geons)
- like how the alphabet can make words, geons create objects
what are the properties of geons?
- vietpoint invariance: can be viewed and interpreted from many different perspectives
- discriminability: can tell one geon apart from another regardless of the viewing angle
- robust to noise: can be perceived even when partially obscured/occluded
what are the short-comings of the recognition by components theory?
- some real-world objects are too complex to break down into simple geons
- need more than structural description to identify things (like size, color, etc)
- sometimes need additional info to determine fine-grained category differences - cannot account for top-down/contextual influences on object recognition
