Lecture Set 3 Flashcards
What does depth perception depend on?
That there is binocular disparity
What is binocular disparity
difference between one eye and the other in terms of things in the world = retinal disparity
How can the stereo correspondence problem be solved?
In order to measure difference in distance - must choose one object on either side
E.g. police man’s hat - match and compare object’s positions
How can the stereo correspondence problem be solved if there are no objects to match>?
Can be done by matching light
What is stereopsis?
First thing you have to do is choose a part on either side of the image to match
- if you correctly match an image on the left and right eye –> calculate distance rom eye
What is an auto-stereogram?
Depth illusion
“magic eyes” - giving yourself a stereogram with one single image
- always have repetition in them –> invites visual system to make a mistake (think they’re closer or further than they actually are)
E.g. with wall paper - repetitions in image –> easy to make mistake
What is retinal disparity?
difference in position (left and right eye)
What is a horopter?
curved plane a distance from person’s vision
What are three outcomes associated with retinal disparity?
- items on horopter
- closer to horopter
- past horopter
What occurs when an item is on the horopter?
Fixation plane - object will lay on exact same position on left and right eye = zero disparity
- no difference between eyes
What occurs when an item is closer to the horopter?
Object falls on opposite side of the eye
- reference point in the middle (same position on left/right retina)
= crossed disparity - object in front of horopter (left side on right eye, right side on left eye)
What occurs when an item is further than the horopter?
Reference point = same position on left and right retina
= Uncrossed disparity - object is on same side of eye but not on the exact same position
Improvements with accuracy and speed with binocular vision?
30% more accurate and 30x more faster (but only at a range of 30m)
What is Panum’s Fusion Area?
area where you can fuse left and right image –> 3D depth image
- if something is too close or too far away from horopter –> double images
Explain the tilt after effect
adapt to some tilted lines, then show some vertical lines –> see them tilting in the opposite direction of those they adapted to
What does it mean when you say the tilt after effect crosses sides?
When you look at slanted lines with just your right and then vertical lines with just your left –> still see after effect (80% of cells respond to information from either eye)
What is a critical period
Experiences early on in life which will play an important role on development (4 weeks to 4 months in cats)
Explain the pirate kitty experiment
patch one eye on (alternating days)
- only getting information from the open eye until critical period ends –> never see things from both eyes at the same time
- retinal binocular cells never develop –> struggle with depth perception (stairs, jumping off things)
What is strabismus?
Imbalance of the eye muscles
What is wall-eyed?
both eyes are pointing outward
- need to be corrected early for binocular cell development
What does it mean to be cross eyed?
Both eyes pointing inward
- left and right eyes see different things so the brain has to suppress one to avoid double images
- never get the same info from both eyes –> requires surgery in the first 4-5 years of life
What is the result of strabismus if it isn’t resolved?
balance problems, can’t see 3D –> auto-stereogram illusion
What is amblyopia?
Cross- or wall-eyed affecting only one eye
- too big a difference between eyes –> brain can’t bring images into coordination so it ignores the eye with the bad image
- cells need to develop before the age of 6 or they will never develop at all
What is binocular rivalry?
when you have two different images in the left and right eye that can’t be matched
- brain typically suppresses one
What is size constancy?
We see objects being the same size despite the changing angle (smaller retinal image)
- same actual size can produce different retinal images
- achieved by visual system making use of all the monocular, binocular, pictorial, motion-based, physiological cues
What does it mean by misplaced size constancy?
Something further away appears larger = Ponzo Illusion
- depth cues are telling us its further away (pictorial cues) so we factor in depth when we interpret how far something is
Describe the moon illusion
moon appears larger when it is closer to the horizon and smaller when it is high in the sky
- but the moon never changes size
- linear perspective, texture gradient, atmospheric perspective, but none when you look up in sky
What is Emert’s Law?
size of after image changes with the distance away of the surface projected on
- burned out area on retina doesn’t change, but visual system is taking into account depth information (when trying to interpret size of after image)
Explain Ames room
Look into room using only one eye (only works with monocular cues)
- constructed in a weird way so linear perspective doesn’t work properly
- personal appears to be changing size when they cross room
- due to lack of binocular depth cues
What is shape constancy?
the shape of an image with vary depending on the relative position of image
- factor in depth
E.g. the dime in a box - treats as a 3D image because of cues
- the dime doesn’t actually fit = shape constancy correction
What do both size and shape constancy require?
Adequate depth information for them to work
- take depth away –> illusions
What is object perception?
ability to see objects as wholes
- recognizing and visual coordination
- happens so quickly (we don’t even realize)
- organizing information into an image
- difficult because objects can be made up of area that are quite far apart –> pointillistic representation
- learning from experience (familiarity)
- combination of perception and memory
- vision –> representation of object –> store as long term memory
What is pointillistic representation?
Each rod/cone knows only about a tiny point of light (not the while thing)
What is the top down theory?
Using long term memory (past experiences) to help determine what you perceived
- use context (e.g. bunny or duck?)
What does perceptual set mean?
see what you are prepared to see
E.g. just watched a horror movie –> see ghosts everywhere you look
What’s a disadvantage of the top down theory?
You see only what you expect to see –> may miss important information
“smart bombs”
What is bottom up theory?
building information straight from retina –> gradually built from pointillistic representation
What stages make up the 3 stage model?
- low level vision (pre-attentive analysis) = bottom up –> spatial parallel
- visual routines (analyses requiring spatial attention) –> spatial serial
- Visual cognition (object recognition) = top down –> space no longer relevant (memory matching)
Describe low level vision
- everywhere in visual field (calculations occurring all over image) all at once
- happens automatically
- categorizing by feature analysis(feature = basic element of an image - line orientation, curvature, colour)
- finding the discontinuities in an image - brightness or colour (edges)
- visual system relies on edges
What does ganzfeld mean?
featureless - no edges = no differences in brightness and colour
- all you see is uniform white
- begin to hallucinate flashes of light
- eyes begin to cross and uncross but you can’t tell because you are unaware of where your eyes are
How do you know an attribute is a feature?
- determined early on
- visual search task
Describe a visual search task
Have a person look at an image to see if a target is there (distractors are also present)
- hit target present as fast as they can it they see target and target absent key if they don’t
This takes time regardless of the amount of items present
What are the Gestalt grouping principles?
Proximity, similarity, continuation, closure
Explain proximity grouping
more likely to group things that are closer together
Explain similarity grouping
more likely to group things that have similar features
Explain continuation grouping
basis to see continuous lines
Explain closure grouping
basis to see closed figures
Describe the masking study
Target = black dot Mask = black donut with a white centre the same size as the black dot
- Little black dot is there for 20ms and then nothing shown after –> you can tell you’ve seen it
- if instead the mask appears after you’ve looked at the little black dot for 20ms –> see the big black dot (visual system is putting images together)
- if you wait 100-200ms before presenting mask –> you just see the mask (replacing little black dot)
- if you wait 200+ ms before presenting mask –> you now see both properly (black dot and mask)
What does the masking study tell us?
Registering edges takes time!
What are some side effects of grouping?
testure segregation and illusory contours
What is texture segregation?
background texture falls apart from different textures (diagonal and horizontal lines)
- only happens in some situations (e.g. T and L + 2 and 10) - spatial relation between parts differentiates them
- T and slanted T differ in line orientation –> seen as separate
- not enough to be in different categories, have to have different feature
What is spatial relation?
on top, to the left, inside/outside, connected
Explain illusory contours
Formation of illusory contours around objects
- Kanisza figure - visual system groups by continuation to see imaginary line to form a cube (with abrupt discontinuation)
- visual system likes to see straight lines
Explain figure/ground segmentation
Determining what’s the background and what’s the object in a scene
- organizational decision (how you organize info)
- people’s interpretation will determine whether they recognize it or not (stored in long term memory)
figure/ground segmentation and spatial frequency
bias to see things with high spatial frequency as the figure
figure/ground segmentation and size
- small = more likely to be image
- large = more likely to be background
Visual routines: analyses requiring spatial attention is used to?
create an object file
Describe visual routines
- one area at a time (like a spotlight)
- trying to find a face in a crowd –> see object as a whole (takes using that one area at a time)
- E.g. Where’s Waldo
What is a feature analysis
E.g. b = vertical line and a circle to the right; blue
- helps differentiate between b, d, q, p (takes time)
How do you know something is spatially serial?
Visual search task - present or absent?
- takes more time, the more items there are in the display
Problems that can occur with visual routines?
- illusory conjunction
- integrative agnosia
- simultagnosia (Balint’s syndrome)
What is illusory conjunction?
Attributes combined - bringing features of one object together with another
- attention needs time to work
- happen more when you’re tired - seeing things that aren’t actually there
- this is why eye witness accounts are often quite inaccurate = “gun focus”
What is integrative agnosia?
Brain damage in temporal area
- problem being able to recognize what you see (can see all the parts just can’t put them together = make object file)
What is simultagnosia or Balint’s syndrome?
- occurs at parietal temporal junction
- can see on thing but not two
- as long as there is no competition can make object file
Explain the final stage: visual cognition
- space no longer relevant (memory operation)
- top down processing
- match object file with one from long term memory
- some people who can’t recognize what they see, can’t make object file –> can’t match with memory representation
What is motion perception?
ability to see things in motion (change)
When is motion necessary for object perception?
Motion can define form = grouping by common fate
- things that move together appear to be part of the same object
- allows you to form representation of an object
- e.g. prey freeze so they aren’t seen
Biological motion: Johansson Figures
- light emitting diodes on joints
- can tell the difference between a man and women just by the way they move
- see thing moving in 3D (sex, species, mood)
What is looming?
Object getting closer appears larger (e.g. baseball towards face)
What is optic flow?
Things changing position on retina
- use of motion signals to let us know when were falling (balance)
- visual capture > vestibular: balance
Describe the swinging room experiment
walls suspended from chains (hanging straight down)
- swing towards baby –> think they’re falling forward so they lean back
- swing away from baby –> think they’re falling back so they lean forward
What is visual capture?
vision overcoming all other cues (e.g. muscular, hearing)
What is real motion?
looking at visual field and things are really moving
- we can’t tell the difference between real motion and the motion created by rapid sequences (apparent motion)
e. g. TV
Explain Gestalt Wertheimer Exner
time between frame 1 and 2 determine what you see
- 1 dot in the corner of one frame and 1 dot in another corner of another frame (time between = 30ms) –> we see on picture with dots on either corner
- if delay between is too short –> static picture with both pictures
- if time between 60-200ms –> see motion
- if time between is 30-60 –> gradual movement
- if time is 200+ ms –> two frames as static images
Two motion processing systems?
- image retina system
2. eye-head system
Explain the image retina system
- eyes still, but objects have to change position on retina
What system corresponds to the waterfall illusion?
image retina
- stare at falls for some time –> look at static image –> appears to be going upwards (opposite direction)
- fatigue cells responsive to downward motion
- cells responsive to upward motion getting least amount of stimulation
- will not work if you try and track water with eyes (moving eyes)
Explain the eye-head system
- see things as moving even when moving eyes
- when we track a object (on fovea) still see motion (e.g. following a rolling golf ball)
- doesn’t necessarily change positions on retina
What is the Aubert-Fleischl Effect
Two systems get motion information in different ways:
- take rapidly moving pendulum –> follow with eyes (back and forth) –> object appears to be moving slower than it really is (using eye muscles to figure out how quickly something is moving) - muscles can’t quite keep up
- take same pendulum –> keep eyes still and watch pendulum = more accurate of actual speed when using image retina system - looks at position from one point to another as it moves on the retina
Explain the type of cell that allows you to determine which direction things are moving
Direction sensitivity
- neurons sending excitatory signal to intermediate cells –> send inhibitor signals to neighbouring neuron intermediate
- positive response with leftward motion (A to F)
- no response with rightward motion (F to A)
Where does motion processing begin?
retinal ganglion cells
What main structures are involved in motion processing?
magno fibers, superior colliculus, lateral geniculate and striate cortex, medial temporal cortex, medial superior temporal cortex, superior temporal area
Describe magno fibers
- big
- thick axon to deliver motion information to the brain
- motion sensitive - transient response
- less common than Parvo but more common in periphery
What is a transient response?
responds best to things that are changing
What is the role of superior colliculus?
- contains the tecto-pulvinar system = 10% fibres from optic nerve (ALL magno)
- motion perception
- cells respond based on whether you are moving eyes
What system is the superior colliculus involved in?
eye-head - controls extra-ocular muscles causing eyes to move
What is the role of the LGN/striate cortex?
- 90% off cells
- emotion perception, complex/hyper-complex cells = respond differently to things that are moving
What system is the LGN/striate cortex involved in?
image retina
- waterfall ills is based on motion sensitive cells of the striate cortex
Explain the kitty in the disco study
Critical period = 4 weeks to 4 months
- certain visual experiences for normal vision development
- stroboscopic displays at discos –> flashes every >200 ms –> complex/hyper-complex cells never develop (appear as static images)
- important to have real motion (where time between 2 images in between 60 and 200 ms)
What is the role of the medial temporal cortex?
Damage –> frozen images
- cells that respond more, the more visual field is moving together
- no correlation between movement - cell responds more with consistency (1/2 cells moving in same direction vs. all cells moving in same direction)