Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
55 Cards in this Set
- Front
- Back
Field of View (Modes of Attention) |
- Horizontal field: 90 degrees on both sides = 180 degrees - Vertical field: slightly less (occluded by forehead/cheek bones) Lecture 9/29 |
|
Binocular/Monocular Vision |
Binocular vision: both eyes can see - Predators have more, for better acuity Monocular vision: only one eye can see - Prey have more, for better FOV Lecture 9/29 |
|
Foveated Vision |
Good detail where looking, poor detail in periphery. Rare trait in animals. Lecture 9/29 |
|
Point Vision |
Uniform density of photoreceptors. Each area of vision can see same detail. Lecture 9/29 |
|
Two Fovea |
One is for central vision, another is the temporal area for tracking things on the ground. Typical of birds of prey and hummingbirds. Lecture 9/29 |
|
Visual Streak |
Typical in dogs and cats. Have an elongated streak on the retina that has a high density of photoreceptors. Lecture 9/29 |
|
Saccades (Mode of Attention) |
Rapid, ballistic eye movements. - Takes 30ms - 3x per second - Occur between fixations (300ms) - Saccadic suppression causes temporary blindness during them - 10% of vision time Lecture 9/29 |
|
Smooth Pursuit Eye Movements |
Only primates have this capability. Must fixate on a moving target or else you make saccades. Primates have them because we have hands so we can manipulate objects close to our face. Lecture 9/29 |
|
Visual Search |
Find something in an array of objects. Two types: - Effortless - Effortful Lecture 9/29 |
|
Effortless Search |
# of distractors doesn't matter. Target has a unique feature that makes it stand out from the rest. Preattentive processing is all that is needed. Lecture 9/29 |
|
Effortful Search |
Reaction time increases with # of distractors. Must focus attention. Lecture 9/29 |
|
Primitive Features |
Features that humans attend preattentively. Must be unique to target for effortless search. - lines vs. curves - angles - color - motion Lecture 9/29 |
|
Parts are transposable (Perceptual Organization) |
For a scene, parts don't matter, but the relationships between them do. Lecture 10/1 |
|
Laws of Perceptual Organization (5)
|
1. Law of Figure/Ground
2. Law of Grouping 3. Law of Good Gestalts 4. Law of Perceiving Whole Objects 5. Law of Perceiving Constant Properties |
|
Law of Figure/Ground (Law of Perceptual Organization) |
Edges have ambiguity (which object does an edge belong to?) Decisions 1. Edge belongingness 2. edge defines shape of figure 3. depth order (what is in front) 4. Background extension (extension behind figure) Lecture 10/1 |
|
Laws of Grouping (Law of Perceptual Organization) |
- Proximity - Similarity - Good continuation: a contour will appear to be connected if occluded and angles are reasonable - Closure - Common Fate Lecture 10/1 |
|
Law of Proximity |
Things that are close together are grouped together |
|
Law of Similarity
|
Things are grouped by similarity (duh)
|
|
Grouping by good continuation
|
If a contour is continous, you'll percieve it to be one thing. Something is well-aligned if it has good continuation. Cannot have more than 1 inflection.
|
|
Grouping by closure
|
If there's good continuation, the brain will complete the contour. A subjective contour is perceived but not really there.
- Modal: not occluded, seen - Amodal: occluded, not seen |
|
Grouping by common fate
|
If it moves together, its' grouped together
|
|
Law of Good Gestalts (Law of Perceptual Organization)
|
"minimum principle"
The default is symmetry and regularity |
|
Law of Perceiving Whole Object (Law of Perceptual Organization)
|
you perceive occluded objects as a whole |
|
Law of Perceiving Constant Properties (Law of Perceptual Organization)
|
(Nothing in notes) Guess: Constant texture gradient, etc. |
|
Inverse projection problem |
There are infinite possible objects that could produce the same visual angle |
|
Depth cues |
sources of information used to compute depth
- Primary: eyes - Secondary: pictorial - Motion-based: objects & self |
|
Primary Depth Cues |
- Accommodation - Convergence - Stereo |
|
Accommodation (Depth Cue) |
changing the shape of the lens - thick: light bent a lot - thin: light bent little good for close objects, at arm's length but not further |
|
Convergence (Depth Cue)
|
verging the eye on a target near object=larger angle far object= smaller angle since you're looking out good for things that are close ***Include picture of angles |
|
Stereo (Depth Cue) |
depth perception based on image disparity - information: retinal disparity, different images projected at different locations - process: stereopsis, computing depth relations provides some information after 20 ft. but not much |
|
Secondary Cues (Depth Cue) |
1. Perspective 2. Familiarity (size) 3. Shading 4. Occlusion |
|
Perspective (Pictorial Clues(?)) (Depth Cue) |
1. linear perspective 2. horizon ratio 3. texture gradient 4. height in plane |
|
Linear Perspective (Pictorial Cue) |
all parallel lines converge on the horizon at the vanishing point |
|
Horizon Ratio (Pictorial Cue) |
has to be objects on flat ground - size of things relative to eye height - Ground must be flat for this to work |
|
Texture Gradient (Pictorial Cue) |
units of texture more dense at the horizon - objects of equal size have the same texture at the base - can scale objects by looking at feet |
|
Height in Plane (Pictorial Cue) |
higher in plane = higher angle of elevation |
|
Familiarity (size) |
- Familiar size scales the space an object is in - People trump everything else when comparing size |
|
Shading |
Assumption: light comes from above. - "above" is relative to retina |
|
Occlusion |
Things that are closer cover up things behind. - Determines depth order but not depth magnitude. |
|
Motion Cues |
- Kinetic Depth Effect (Structure from motion) - Motion Parallax |
|
Kinetic Depth Effect (Motion cue) |
A solid object rotated gives depth info and defines shape. - Must be rotated on any axis except line of sight. - Solves inverse projection problem |
|
Motion Parallex (Motion Cue) |
The closer something is, the faster it appears to move. - Fixation causes things to move at a rate proportional to how far it is from point of fixation (the closer the less it moves) |
|
Properties Revealed in Motion |
1. Surface Segregation 2. 3D Form - Rigid - Hierarchical
|
|
Surface Segregation |
Solves figure/ground problem when object moves on background. - Dynamic Occlusion: When a surface moves over another, you can't help noticing the revealing/covering of texture. - Most powerful cue - M-Cells detect this motion |
|
Hierarchical 3D Form |
Consists of 3 types of motion: - Absolute - Relative - Common |
|
Absolute Motion |
Movement of individual components |
|
Relative Motion |
Movements relative to one another. |
|
Common Motion |
Motion as a unit. |
|
Focus of Expansion |
How we are able to tell where we are going. Texture moves away from stationary point of focus. Measured by: - Time to Contact - Time to Passage |
|
Time to Contact |
Used to determine how long until you collide with something. Visual angle divided by change in angle per second. (tau) |
|
Time to Passage |
Used to determine how long until something passes you. Visual angle divided by change in angle per second. (tau) |
|
Linear Optical Trajectory Heuristic |
Position yourself so that trajectory is linear so that you will intercept the path of the airborne object. Not optimal because it causes a curved travel path, and doesn't specify where the landing spot is. Also doesn't provide environmental info. |
|
Perceptual Development: Color |
- 2 cones (not blue) for newborn - 3 Cones at 3 months |
|
Perceptual Development: Acuity |
- 20/500 at birth - 20/40 at 7 months Babies and adults differ in: - Photoreceptor shape - Density of photoreceptors - Cell connectivity Babies have short, wide receptors that are less dense in the macula, and less connected than that of adults. |
|
Cell Migration |
The process in which photoreceptor cells in infant eyes move toward central vision. |