Exam 2: Sensation and Perception

contrast borders

contours that distinguish objects from their background

contrast borders are detected by...

ganglion cells

local edges

contrast borders that fall within a cells receptive field

texture contrast

distinct varying textures on objects/surfaces (though average intensity may be the same)

texture boundaries

perceptible differences in texture based on size, shape, orientation, color, or brightness

perceptual grouping

automatic grouping of elements that are visually "similar"

Gestalt movement motto

'the whole is greater than the sum of its parts'

apparent motion

principle used in animation, an object appears to move when it rapidly changes position

Law of SIMILARITY

objects that share visual qualities are grouped

law of PROXIMITY

objects that are near each other are grouped

law of GOOD CONTINUATION

contours that flow smoothly are grouped into a single contour

law of COMMON FATE

objects that move in a similar direction or pattern are grouped together

law of CLOSURE

contours that touch or appear to form closed shapes are grouped

why do Gestalt Laws "work"?

because they reflect WORLD REGULARITIES- symmetry, parallelism, good continuation- that have been internalized by evolution

figure-ground perception

automatic distinction between an object and background

principle of non-accident

perceptual interpretations involving coincidental viewpoints/alignment are unlikely

figure side

shaped by the contour (has an outline)

background side

unbounded (does not have an outline)

recognizing contours of negative space...

longer response time, less accurate

why does memory prefer figure ground?

because objects can still be recognized when in motion, but negative space will be less recognizable

figure ground determiner: closure

fully enclosed regions are more likely to be perceived as figure ground

figure ground determiner: size

smaller areas are more likely to be perceived as figure ground

figure ground determiner: convexity

convex contours are more likely to be perceived as figure ground

figure ground determiner: symmetry

symmetric contours are more likely to be seen as figure

the problem of occlusion

most objects are partially hidden behind/covered by others

visual completion

perceived continuity of partially occluded objects

amodal completion

a partially occluded object is completed

modal completion

an object that is partially camouflaged or seen in low illumination is completed

illusory contours

perceived contours and cues to depth that don't physically exist

recognition is different from perception because it involves...

memory, categorization, and naming/labelling

Marr and Biederman theory of object recognition

objects are stored in the brain as 3D models with simple parts called "geons"

Recognition by Components

all objects can be mentally constructed using geons

3 properties of geons

1. shape of cross section

2. size of cross section

3. axis curvature

cross section

shape created by making a cut down the center of an object

pros of geon models

- recognizable from ALMOST any viewpoint

- recognizable under partial occlusion

- a small number of geons can be combined into a large number of objects

weaknesses of geon models

- don't contain a lot of info at finer levels of detail

- accuracy drops when recognizing a geon at a different angle from when it was learned

- doesn't account for subordinate level recognition (ex. dog > breed)

View Based Recognition (VBR) model

objects are stored as the compilation of multiple viewpoints

inferotemporal cortex (IT)

area in the cerebral cortex responsible for object recognition, found via lesion deficit method

neurons in the IT

- respond to complex shapes

- have large receptive fields

-response levels rise when presented with complex stimulus

pros of VBR model

- neural responses are translation invariant, they respond no matter the angle and size of stimulus

- respond to partially occluded objects

Fusiform Face Area (FFA)

brain area responsible for facial recognition, found via lesion deficit method

processing faces vs. objects

faces are processed holistically rather than as a sum of attributes, stored distinctly from other objects

Larry and Larry's House study

participants could recognize individual components of Larry's house (the door, the roof) but not facial features (nose? mouth?)

Contrast Reversal and Facial Recognition

facial recognition declines when light and dark values in an image are inverted

Prosopagnosia

face blindness, inability to recognize familiar faces

Monocular vs Binocular

one eye vs two eyes

pictorial depth cues

monocular depth cues found in a single frame (still image)

texture gradient

variations in texture density, number of texture elements per unit area

shading

varying light and dark values on a surface

shadows

dark values that imply blocked light

ball-in-a-box demo

the trajectory of the shadow in relation to the ball changes perceptions of depth

kinetic depth effect

perceiving 3 dimensionality as a result of object rotation

retinal disparity

the difference in lateral distance between objects in the L and R retinal images

T-junction

where one object "disappears" behind another

transparent X junction

point where four surfaces of different values overlap

contrast polarity

matching direction of contrast implies transparency (ex background is darker on the left so the transparent object is darker on the left)

linear perspective

parallel lines converge as depth increases

one-point perspective

all lines in an image converge to a single vanishing point (they don't have to actually touch)

Ponzo illusion

same-sized objects placed between converging lines will look different

two point perspective

lines converge at two opposing vanishing points

atmospheric perspective

objects become hazier and less detailed as they recede in depth due to scattered light

depth and size

but objects that appear to be father away are interpreted as larger than their fixed size on the retina

afterimage size

result of fatigued photoreceptors, demonstrates that image size is fixed on the retina

corresponding points

two points on the left and right retina that are the same degree of distance from the center

relative disparity

the difference in absolute disparities of objects in a scene

when relative disparity = 0

objects are perceived as being at the same depth

horopter

a curved line that represents objects which have fall on corresponding points and have a relative disparity of 0

stereopsis

binocular cue, 3 dimensionality emerges as a result of retinal disparity

crossed disparity

non-horopter points behind the point of fixation have uncrossed disparity, those in front have crossed disparity

Julesz experimental goal

to disprove the old conception that binocular vision was a result of object recognition between left and right retinal images

random dot stereogram

panels of randomly shaded squares that when viewed through a stereogram produce a 3D image, Julesz

evolutionary significance of color vision

it is evolutionarily advantageous in quickly identifying food sources (ex. red ripe berries)

electromagnetic spectrum

full range of wavelengths of electromagnetic radiation

visible spectrum

the range of visible wavelengths

why is wavelength insufficient in defining color?

- fails to explain why wavelength corresponds with color

- physical wavelength can be different from perceived color (ex in different lighting conditions)

T/F: color is an intrinsic property of light

FALSE

reflectance function

the overall percentage of light that is being reflected

luminance

light intensity that reaches the eye

iso-luminant

differently colored but same light intensity

T/F: color is perceptual rather than a physical property

TRUE

Newton's Prism experiments

white light passed through a prism to produce colored beams of light

white light

the combination of all wavelengths on the electromagnetic spectrum

spectral colors

monochromatic wavelengths

T/F: sunlight is highest energy at 450nm

FALSE- near equal at all wavelengths

complementary colors

two wavelengths that combine to produce achromatic light

T/F: yellow + blue = achromatic light and

green + red = achromatic light

TRUE

subtractive color mixing

non-visible colors are absorbed, all that is left is the new color

additive color mixing

all present wavelengths are combined

mixing colored paints and pigments

subtractive

mixing colored lights

additive

metamers

mixtures of wavelengths that are physically different but are perceptually identical

univariance principle

cone types do not contain any information about color, only about amount of light being absorbed

absorption rate

percent of light absorbed, in photons

how does absorption rate explain metamers?

the brain cannot differentiate monochromatic light vs combined lights when the net amount of light is the same

L cones respond to

red wavelengths

M cones respond to

green wavelengths

S cones respond to

blue wavelengths

an absorption rate of 50% translates to ?/400 photons

200

dichromacy

only possessing two cone types

trichromacy

posessing all three cone types