LECTURE 3 - Perception Flashcards
what is perception?
- ability to recognize and interpret information from the senses
- provides building blocks for our thinking and remembering
what is mental imagery?
- we can imagine an object without seeing/hearing/touching
- this is made possible by perception
what are some examples of ways that our brains fill in missing information?
- eyes fill in the blindspot from the retina for cohesive visions
- visual illusions take advantage of how our brains fill in information
how does the blue/black, gold/white dress show us that perception is an active process?
- some people see one colour combination, some see the other, and some can switch between the two
- difference in perception is due to differences in our minds’ assumptions about lighting conditions
what are bottom up processes?
- sensory input from receptors to the primary sensory cortex for that system
- input leads to sensation, detection of sensory signals and the beginning of perception
- transduction: conversion of physical stimuli into electrochemical signals used in neurons
how does an image coming into our eye differ from reality?
when an image enters the eye, it becomes upside down and inverted
where on the visual field are the hemispheres responsible for vision?
- both eyes get input from both sides of the visual field
- axons from the nasal half of each eye cross over to the opposite side
- each hemisphere receives input from the opposite visual field, but from both eyes
what is the primary visual cortex? who discovered it?
- also called V1
- david hubel and torsten wiesel
- specialized for basic visual processes; orientation, spatial frequency, motion direction of light and dark patches
after V1, how does visual information get processed?
- sent forward in the brain via feedforward connections to serve higher order functions
- perception of colour, motion, objects, faces
- part of bottom up processing
- areas in frontal, temporal, and parietal cortices interpret information to guide thought and action
what are top down processes?
- knowledge and expectations that influence and enhance out interpretation of sensory input
- feedback connections resolve ambiguity of feedforward signals (bottom up processes) that are subject to multiple interpretations
- goes from more sophisticated brain areas to more basic areas to change how we view a stimuli
what are some things that affect the way we perceive stimuli?
- context
- past experiences
- brain makes predictions
what is an example of how context is top down processing?
- context shapes perception and facilitates object recognition
- when you read the image from left to right its ABC, but without context it could be A - 13 - C
what is unconscious inference?
educated guesses about most likely interpretation based on visual clues
- coined by H v. Helmholtz
what is the theory of predictive coding?
- the brain constantly generates predictions about the world
- expectations are constantly compared with and validated against incoming bottom-up signals
- predictions and expectations guide the mind to the most likely interpretation
what is object segmentation?
- visually assigning the elements of a scene to separate objects and backgrounds
- hard problem for the visual system to solve
- bottom up cues like visual motion can help with this
- but bottom up cues can be insufficient for a computer to distinguish objects, suggesting we need top down processes as well
what is figure-ground organization?
- a form of object segmentation where boundaries of objects are clearly visible, but
- it is unclear which side belongs to an object and which side belongs to the background
what rules does the brain use to solve ambiguity in figure-ground organization?
rules for determining what the object is
- rule of enclosure: region surrounded by another
- rule of symmetry: more symmetrical region
- rule of convexity: bulging region (convex)
- rule of meaningfulness: meaningful region
- if we can assign semantic meaning to something, we are likely to label it as a figure
how do we see things despite them being partially occluded by other objects?
- the brain fills in missing info, also called amodal completion
- we are able to follow moving objects when they are occluded by extrapolating their trajectory based on their motion path
what is the inverse projection problem?
- 2D input has to be converted to a 3D representation
- multiple 3D representations are possible from a single 2D image
- top-down information (depth cues) is needed to disambiguate the possibilities
how do depth cues help us see in 3D?
- binocular disparity: retinas of the two eyes receive slightly different images, the closer something is to you, the greater the difference between what your two eyes see
- binocular depth cues: calculate how close an object is to us based on the extent of disparity between the eyes
- monocular depth cues with one eye (pictorial cues) also help us see 3D; linear perspective, texture gradient, blur, shadow/shading)
what other cues do observers use to solve the inverse projection problem?
- size constancy: we perceive an object as the same size even when its distance to us changes
- we use our experience to evaluate object size, then use object size to determine distance - motion parallax: the movement seen when the observer moves through a stationary environment
- the closer an object is, the faster it seems
what is relative motion?
when stationary objects appear to move
what is colour constancy?
our visual system factors in differences in illumination when perceiving colour so that we perceive the same colour regardless of illumination
what is lightness constancy?
we factor in illumination conditions when perceiving the brightness of things
what is colour assimilation?
- the opposite of colour constancy
- the color of an area appears to take on the hue of nearby colours
what is object recognition and how can it be challenging?
- ability to match images to stored representations of them in the brain
- difficult because of variations in shape, orientation, and lighting conditions
- trying to decide if two images are of the same thing depends on familiarity
what is agnosia? what can it help researchers understand?
- the inability to recognize objects and to match them to correct categories and labels
- allows researchers to distinguish between early and late visual processes that contribute to object perception
what is apperceptive agnosia?
impaired early vision
- people cannot perform simplest visual feature task
- can’t discriminate orientation, colour or motion
- suggests problems with bottom-up processes
what is associative agnosia?
impaired late vision
- people cannot name or categorize objects
- suggests difficulties with top-down processes
what is the view-based approach to enabling object recognition?
- we match images to representations that are like 2D pictures or templates
- a template is a representation that filly describes the shape of an object
- simple and easy for clearly defined objects like numbers
- can be done by machines
- difficult for natural images and real-world objects
what is the structural descriptions approach to enabling object recognition?
- object recognition is based on models that represent objects as sets of 3D parts standing in spatial relationships to each other
- models have a “vocabulary” of components that are combined to form an object
what are some pros and cons of the structural descriptions approach?
cons: slow, hard to learn, unnatural
pros: flexible, objects identified regardless of orientation, angle, illumination, etc
- more tolerant of image/object noise than view-based approach
do we process objects and faces holistically? meaning as a whole object at once
- experience helps shape the degree to which object recognition is holistic
- experts in certain areas tend to process associated objects holistically
- we are all face experts, so we process faces holistically
what research was done on how experts view objects holistically?
- all people had a specialized area for processing faces (fusiform area)
- face areas were also active when images of the area of expertise were shown
- FFA responds to non-face stimuli if they fall into the area of expertise
who is yasmin el-shamayleh and what did she discover?
- one of few racialized women doing research on primate object recognition using single-cell recordings and optogenetics
- found that each neuron represents a small part of an object and many neurons must work together to represent the whole object
- findings support the structural description approach to object recognition
what is mental rotation?
a type of mental imagery where we are able to compare and match rotated images
what is aphantasia?
condition where a person has an inability to form images in their mind
what are the two pathways for visual information after the primary visual cortex (V1) ?
- dorsal pathway to posterior parietal cortex
- ventral pathway to inferotemporal cortex
what is the dorsal pathway of vision responsible for?
“where” pathway, action pathway
- spatial orientation
- motion of objects
- attention
- uses perceptual info to guide ongoing actions
what is the ventral pathway of vision responsible for?
“what” pathway, perception pathway
- object recognition and colour
- supports many perceptual functions
what is the updated duplex theory of vision?
- visual system has two separate interacting components
1. one for perceiving object
2. one for controlling actions directed at the objects - both streams use info for distinct purposes and analyze visual information separately
- they work together to produce adaptive behaviour
