2 Flashcards
(53 cards)
perception
study of how external world is represented in our brain or mind so we can understand and act on what is around us
agnosia
type of condition where patients have a visual perception deficit due to cortical damage, despite normal vision
apperceptive agnosia
unable to name, match, or discriminate visually presented objects. can notice basic visual features (things that make up stuff like lines and color), but these traits are disconnected and cannot copy or draw them correctly. cannot match, name or discriminate. cannot combine basic visual stimuli into a complete percept
associative agnosia
patients cannot associate visual pattern with meaning. are able to combine visual features into a whole, so they are able to reproduce images. but do not know meaning so tend to reproduce with more detail than an average person. because intelligence and memory is intact, they can guess what an image is
how can patients with associative agnosia form understandings of how objects work
pick up an object and suddenly they can use them. may then realize what object is
agnosia patients tell us there are three dissociatable steps to visual perception, which are:
input/sensation (removed by blindness), basic visual components are assembled (apperceptive agnosia cannot do this and cannot form a representation). meaning is linked to visual input (associative amnesia is fucked up here)
experience error
false assumption that structure of world is given directly from senses (representation doesn’t always represent world; visual illusions illustrate that we don’t always percieve accurate representations)
eyes recieve input from what part of the fixation-saccade cycle
the fixation phase
what is a jumping saccade
when you are moving eyes around, basically eye movement that is not smooth pursuit
smooth pursuit
tracking object that is moving
computational approach to studying perception
how the brain and interprets distal stimulus (i.e. stimulus that is away from you in the environment)
proximal stimulus
a distal stimulus when picked up by a receptor
perception/action approach to studying perception
goals of action determine perception. action is more important than understanding
gestalt approach to studying perception
organizational principles to create meaningful perceptions of the environments, bringing parts into a whole
bottom up (data driven) processing
converge info drom input. it can be used purposefully
top down (conceptually driven) processing
perception is influenced by prior knowledge, memories, and experiences. automatically used, and needs input, but the memory is fed forward to speed up processing. expectations influence perception. for example you automatically look for blonde people if looking for a blonde friend.
template matching
we have a mental stencil for different patterns. does not work for complex stimuli and originally for computers. works for things like barcodes
feature matching
pandemonium theory by selfridge. states that we have a system for analyzing every distinct feature of a visual item. image demons provide input to feature demons, which give output to cognitive demons, which yell to the decision demon
is pandemonium theory an example of serial or parallel processing
both. demons counting stuff at the same time
physiological support for feature matching theory
there are feature detector neurons in V1 that respond to different line orientations (vertical line detector) V1 has all orientations
biederman’s recognition by components aka geon theory. what was it used for
theory of human vision to be applied to computers. need viewpoint invariance for computers. involves basic 3D shapes. to recognize these shapes, objects are broken into geons.
geons are viewpoint ___? issues with this?
viewpoint invariant because they have nonaccidental properties (not due to viewing angle). biederman says that as long as you can pick up geons, you can understand a perspective. however, recognition is impaired when you view objects from non-canonical viewpoints
is it still perception if you look at a noncanonical angle and need time to recognize the object
no
does recognition by parts discuss how it is physically implemented in humans
no, it comes from a history of computational theories that are not concerned with how it works in humans