Final Flashcards
Where does the visual system break things apart?
Point of concavity
Advantages of Elastic Graph Representation
- Invariant to position, size, orientation in plane (2D) and some degree of view-point
- Biologically plausible
- Nodes can be represented as hypercolumns in the brain and connections as synapses
- Accounts well for holistic and subordinate recognition
What makes Elastic Graph Representation different from template?
Representations are encoded in addition to local features
Feature-set representation characteristics
- Ex: Mel’s Seemore
- Parallel neurons fire for specific features and when the right combination of feature neurons fire, the object is identified
-Fast because of parallel processing
Potential disadvantage with feature-set representation and Seemore’s reply
- Visual system would erroneously identify an object if the features were scrambled
- Possibly visual system could require specific features that would prevent scrambling
Why is motion perception so important?
Retinal images are always moving even in a still scene
- Humans move
- Objects in retina move
- Micro saccades and deliberate saccades
Subject-relative motion
Word for when the only thing moving is the subject?
What is the slowest an object can move and still be perceived?
10-20 min (1/60 of a degree) per second w/o background (subject relative)
1-2 min per second w/ background (object relative)
Retina displacement
Changing position of an object position’s in your retina
optical pursuit
Term for when an object stays steady in our fovea and we track it with our eyes
Apparent motion, beta motion, stoboscopic motion
Perceiving an object when their is no motion (TV)
Interstimulus interval
Term for time between the end of one flash and the beginning of the next
Interstimulus interval thresholds
If ISI < 60 ms we perceive moving object as being in two places at once
If 60 > ISI < 200 ms we perceive fluid motion
if ISI > 200 ms we perceive succession
One part of the visual field or an object makes another part or the entire visual field of the viewer appear to move
Induced motion
Indirect Perception Theory
Motion perception depends on other processes in the brain
Direct Perception Theory
Motion perception is directly computed from inputs
How much time needs to pass between two stimuli for people to determine the presentation order?
> =45ms
How much time needs to pass between two stimuli for people to tell which direction the same stimulus has moved?
14ms
What was the conclusion of the Exner study and how did they reach that conclusion?
They found that people could tell motion direction at much smaller intervals than presentation order.
-Direct perception theory must be true
What was the conclusion of the Wertheimer study and how did they reach that conclusion?
We can perceive motion without perceiving an object (objectless motion)
-Ganglion cells (early in the visual system can perceive motion)
When IST is low and an object appears to be in two places at the same time
Phi motion
How does a Reichardt detector work?
- Neuron A and B have adjacent perceptive fields
- Neuron A synapses with D causing a delay
- Neuron X only fires when it receives simultaneous input from D and B
How are Reichardt detectors arranged to detect the complexity of motion in the visual field?
Reichardt detectors are arranged in opponency with right detector inhibitory and left excitatory or vise versa
What is the waterfall illusion and how is it caused?
If you stare at a waterfall and then avert your gaze to the surrounding rocks, you will see the rocks as moving upwards because downward reichardt system becomes fatigued
What is the stepping feet illusion and why does it occur?
- When a yellow block and a blue block move across a black and white stripped field at the same speed, they appear to move one at a time
- Motion is perceived by the magnocellular-ventral stream which is color blind
- When the yellow block passes over the white stripe it cannot be perceived because there is not enough contrast and same thing occurs with the blue block and the black stripe
- Consequently, the stripes seem to move out of phase with each, one at a time
- If you turn down the contrast on the background, the blocks appear to move synchronously and smoothly
term for change in position
First order motion
Term for change in change
Ex. in noisy square, black and white squares change color
Second order motion
-DIfferent mechanisms than first order motion
Two main problems when computing global motion from local motion detectors
Correspondence problem - identifying which points belong to the same object as they appear and disappear across the visual field
Aperture problem - When viewed through an aperature (receptive field), the movement of a local feature is ambiguous
Where in the brain is motion primarily processed after V1?
Medial temporal and medial superior temporal area or V5
Why does the world not shake when we move?
Motor signals stimulate corollary discharge signals that reach a comparador and then reach the visual cortex
The geometric term describing motion of an object in space during which there are no changes in distances between any two points on the object. (object does not get distorted as it moves)
Rigid Motion
Can the brain deal with non-rigid motion?
Yes, ill-posed motion. Elastic motion, joint motion
What is the difference between depth and distance?
Depth is the relative position of an object from the observer (nearer vs. farther).
Distance is absolute (how many arm lengths etc.)
What is the difference between metric and non-metric?
metric is qualitative
Motion parallax
When an observer moves, objects in the visual field move. The extent to which the objects move indicate how far they are from the observer.
Optic flow
Movement of the whole field (motion parallax)
What algorithm does the brain use to divine edges from motion?
Nonzero differences in optic flow values between an object and its neighbor indicate an edge.
Cilliary muscles
Muscles that manipulate the width of the lens
What is accommodation?
Ability of eyes to change lens shape to focus on objects at different distances
What do ciliary muscles do when the eye focuses on a distant object?
Cilliary muscles relax causing lens attachment fibers to become taut and flatten lens
What do ciliary muscles do when the eye focuses on a close object?
Cilliary muscles relax contract releasing tension on fibers and allowing lens to become more spherical
What is convergence?
Eyes converge towards each other when an object is close
How does the brain know how far apart the eyes are?
Proprioceptive information
What are oculomotor cues useful for?
Calculating how far objects are within in near space (1 - 3m) from us
Pictorial cues
Depth cues that can operate in flat pictures. All of them are monocular cues in that they can operate when you view with only one eye
Laws of optics
How a 3d world is projected on to a 2D retinal image
-Laws of physics that the perceptual system is implicitly aware of
Inverse optics
The brain reverse engineers the projection of the 3D world on to the retina
-Ill-posed problem
Law of linear perspective
- Brain judges on object further down a set of converging lines to be longer because it things the object must be farther away.
- Parallel lines in the image plane remain parallel in other planes they converge
Texture Gradients
Areas that have greater texture density (more, smaller units in the same space) are perceived to be farther away
-Metric
Familiar size cue
The brain keeps track of how big an object is
-Metric
Relative size
Novel objects that look similar but are different sizes give clues to depth in a metric way (this object is half as large so twice as far away)
-relative metric
Relative height
Objects higher in the visual plane will appear to be deeper in 3D space
-relative metric
aerial perspective cue
Objects that are farther from us tend to appear lower contrast and more bluish
Is occlusion metric or non-metric?
non-metric
Is binocular disparity metric or non-metric?
relative metric
reliability - mathematical definition
inverse of variability
What is the difference between the accidental and generic view
- The generic view is the view that allows you to model the object in the same way as you would from most other perspectives.
- The accidental view is a unique model that can only be generated from one particular view
Who was the first scientist? (who invented the scientific method)
Al-Haytham/Alhazen
Who was the father of experimental psychology and the father of optics?
Al-Haytham/Alhazen
How is unconscious inference made?
The brain computes every possible interpretation of visual data and uses prior information to estimate the probability that it matches reality. Then, it chooses the interpretation that has the highest probability of being true
What is Bayes rule?
P ( S X|I )∝ P (I|S X) P ( SX)
prior likelihood posterior
Optical explansion/contraction
When an object moves towards us it expands in our visual field. Symmetrical expansion means its on a trajectory to hit us.
Accretion/deletion of texture
Occluded objects are farther away than occluding objects. As occluding objects move, they will accrede (show) or delete (hide) the objects behind.
Steroscopic info
Info we get from using both eyes
Binocular disparity
Differences in two eye’s view of the world. More difference = greater depth of object in visual field
Where in the visual field do object project on to the same point in both eyes
fovea
Crossed disparity
Indicates that a point is nearer to the observer than the point being fixated
Uncrossed disparity
Indicates that a point is farther from the observer than the point being fixated
Horopter
A set of points in the world that have identical binocular disparities
How does a random dot sterogram work and what is the significance of its effect?
If one eye sees a box of noise and the other sees a complimentary box of noise, a person can see a square
suddenly emerge
Shows you don’t need other visual info for binocular disparity to work
What is the threshold of human sound perception in hz
(where 1 Hz = 1 cycle/sec of vibration) ?
20 - 20,000 Hz
When does sound perception begin to decline in humans?
After adolescence sound perception begins to get worse, it generally stabilizes in adulthood and continues to decline in old age
What is the threshold of human sound perception in decibels?
-10 to 150, high risk after 115 and 150+ pain
What is a interaural time difference?
Difference in space between the two ears allows auditory system to locate the sources of sounds in the environment
What is an azimuth?
An imaginary circle that extends around us in a horizontal plane
What is the maximum interaural time difference and where would it occur?
It would occur to one side or the other of you
What spatial information can be extracted from phase differences and when can that be extracted?
Differences in the phase of a soundwave arriving at two ears can be used for localization but not when the frequency is high
What is interaural intensity difference?
The difference at level (intensity) between a sound arriving at one ear versus the other
How do phase differences and intensity differences compliment one another to allow for localization of sound?
1) Phase differences work for low frequency sounds.
2) Intensity differences work for high frequency sounds.
3) The combination allows us to localize most sounds.
Cone of confusion
The cone of confusion refers to a set of points in space that produce identical onset, phase or intensity differences, due to symmetries of being in front / behind the head, or above / below the head
How can the cone of confusion be resolved?
- Head movement
- The pinnae funnel channels certain sound frequencies better than others, and the intensity of each frequency varies depending on the direction of the sound.
What is inverse acoustics?
When the brain uses what it knows about the shape of the pinnae to reverse the laws of acoustics
How does the brain use frequency information to figure out how far an object is?
High frequencies are interpreted as close because most far-away objects are low frequency
What are the four biggest clues the brain picks up on to determine how close an object is?
– Simplest cue: Relative intensity of sound – Inverse-square law: As distance from a source increases, intensity decreases faster such that decrease in intensity is distance squared
– Spectral composition of sounds: Higher frequencies decrease in energy more than lower frequencies as sound waves travel from source to one ear
– Relative amounts of direct vs. reverberant energy
(If you are farther away from a sound source, the sound will have had more space to bounce off objects)
What is the fundamental frequency (harmonics) and what happens if we take that frequency out?
The fundamental frequency is the lowest frequency and if you take it out, people still here the lowest frequency (called “missing frequency effect”)
What accounts for the fact that people here a missing fundamental frequency?
Since all frequencies are multiple of the fundamental frequency, the wave pattern still repeats every period of the original frequency.
What is timbre?
Associated with harmonics of sound. Quality that distinguishes the sound of a violin playing a melody and the sound a piano playing that same melody
What is sound onset and why is it important?
- How smooth or abrupt a sound is
- Auditory system is fined tuned to pick up small differences in onset
- Ex. ‘ba’ vs ‘wa’
- Ex. C# note played on violin plucked vs played with bow
What is the cocktail party problem?
How do you segregate, identify and process particular sound sources?
What are the three main sound source principals the brain uses to solve the cocktail problem?
– Spatial separation between sounds
– Separation on basis of sounds’ spectral or temporal qualities
– Auditory stream segregation: Perceptual organization of a complex acoustic signal into separate auditory events for which each stream is heard as a separate event
How does the brain group sounds?
- Sounds that have similar frequency are grouped together
- Sounds that pop out by not following the same pattern as the rest of the sounds in a sequence are grouped separately
- Sounds are grouped by timbre (Ex. All notes played by violin together and all notes played by a piano)
What would you hear if there was a small cough in the middle of a sentence?
The brain would perceive the cough as an occlusion and fill in the gap with the appropriate syllable. The listener would be unaware of any interruption or abnormality in the sentence
What are the three primary characteristics of the emerging paradigm on sensation in the brain?
- Multisensory
- Non-hierarchical
- Integrative (non-modular)
Why do many lay people believe that vision is the dominant modality?
We rely on vision a lot in the modern world. Electric screens allow us to absorb large amounts of information with our eyes and electric lights allow us to see long into the night
What two experiments made psychologists believe vision was the dominant modality?
- Ventriloquist illusion shows vision is dominant over hearing
- Visual capture shows vision is dominant over proprioception and somato-sensation
- McGurk effect shows vision can influence speech perception
- Pluck and bow effect shows vision can influence auditory perception
What is the ventriloquist illusion and is it common in everyday life?
- Ventriloquist illusion showed that people will attribute sound to a moving mouth even though it makes more sense that the puppeteer is making the sound
- Common illusion: Audio dialogue comes from speakers but we attribute it to actors on screen
What is the visual capture illusion?
When people are given goggles that shows their hand is in a different spot then where it really is, they will report the location of their hand in the goggles not the real location
-Continues even when hand is touched
What is the McGurk Illusion?
Actor mouths symbol “Ga”, audio plays sound “Ba”, we hear “Da”
What is the pluck and bow illusion?
When people saw a cello being plucked, they perceived the bow sound to be more pluck like then when they saw it being bowed (same sound played in both videos)
What is the stream-bounce illusion and what are its implications for visual dominance theory?
When people see two identical balls shoot towards each other, they are more likely to believe the balls hit and bounced off one another when they hear a clanging sound.
- If they do not hear the clanging sound, they tend to thing the balls went through each other
- Illusion demonstrates that auditory stimuli can disambiguate visual stimuli but not taken seriously initially because both streaming and bouncing are valid interpretations of the scene
What is the sound-induced flash illusion and why is it important?
While fixating on a central dot people see a flash paired with a tone on one-side of their peripheral vision and a flash paired with two tones on the other
- When the flash is paired with two tones, they often see two flashes
- This was the first time another sensory modality (hearing) was shown to be dominant over vision
What is the FMRI evidence that sound influences vision?
When a sound is paired with a visual stimulus, activation of visual processing areas including V1 was higher
- Evidence suggests it is not an attentional phenomenon because visual activity was higher when subjects viewed sound-induced flash illusion than a non-illusionary stimulus
- Activity in visual areas from illusionary flashes parallels that from real flashes
What is the underlying neural circuitry behind multisensory perception?
There are extensive projections from primary auditory cortex and parabelt and STP to peripheral V1 area and peripheral V2
What is the difference between accuracy and precision?
Accuracy is how correct you are precision is the level of correct detail you provide
What is the advantage of multisensory integration?
- Increases both accuracy and precision
- Increases reaction time
- Providing more complete info about the world
_______ are the basic units in our representations in the world
objects
Why must the visual system frequently solve the problem of grouping?
Parts of objects are almost always occluded
What is basic level recognition?
Identifying the class of the object
What is contour integration
Connecting multiple edges together to form contours even when an occluding object interrupts the continuity of the edges
What are four major challenges in edge detection?
– Missing edges
– Spurious edges
– Integrating local edges into global contours – Distinguishing different kinds of edges
• Object boundary (depth discontinuities)
• Surface markings
• Shadows
• texture
What are some sensory clues in edge detection?
- luminance
- color
- motion
- depth
- texture
What are non-accidently features?
Features that appear in the retina because they convey some pattern or regularity
What is a “T junction” and what clue does the brain take from it?
A darker object is occluding a lighter surface
What is an x junction?
Occluding surface is transparent and you can still see the occluded surface behind
What is an y junction?
Indicates corner of 3D object
What is a L junction?
Corner of a 2D object
What are the Gestalt principles
- Common Fate
- Similarity
- Proximity
- Good Form or simplicity
- Good continuation
- Closure
- Symmetry
- Figure-ground = an object is a figure or the background it cannot be both
What are problems with Gestalt principles?
- Description but no explanation
- They lack quantitative models
- Do not explain which principle will win if multiple are matched against one another
Contour interpolation follows a smoothness constrain known as contour _______
relatability
Why is shape so mysterious?
Shape is orientation and scale invariant so it cannot be described to be a sum of local features. It is based on the relationship between the local features
What are the advantages of volumetric representations as a theory of object perception and why is it probably not true?
Volumentric representations are view, scale, orientation and pov invariant but the equations to build them are incredibly complex. Probably too complex for the brain
Limitations of geon theory
- Some people question if it is really viewpoint invariant (greebles)
- It is pretty clear that geons are insufficent for making subordinate distinctions