COLOUR - Department of Psychology
Download
Report
Transcript COLOUR - Department of Psychology
Psy280: Perception
Prof. Anderson
Department of Psychology
Vision 6
Colour, depth and size
1
Need for colour
Some tasks are impossible without it
Can you find the word?
2
COLOUR: What's it for?
Identification / discrimination
Detection (non-detection)
Detection
Potential mates, enemies, prey
Camouflage
3
What are colours?
Light varies in both intensity and wavelength
Light of different wavelengths appear as different colours
4
COLOUR: ATTRIBUTES
THIS IS
NOT RED!
It is 690nm
Colours don’t exist – they’re in our heads!
Psychological property
Interaction: physical light - nervous system
There are no color, just wavelengths…
5
Newton’s dorm room
experiment
Light through prism
= rainbow
Why?
Diff wavelengths have diff
refractory properties
Long (red) bent least, short
(blue) most
6
COLOUR: ATTRIBUTES
Isaac Newton (1666): “colour” of light.
White light (sunlight) = sum of components
Individual component = different colour exp.
Colour = wavelengths subtracted from light
7
Redux:
Do wavelengths have colour?
“The Rays to speak properly are not coloured. In
them there is nothing else than a certain Power and
Disposition to stir up a Sensation of this or that
Colour…” Newton
Different sensory system would result in different
rainbow
8
SHORT
400-450nm violet
450-490nm blue
MEDIUM
LONG
500-575nm green 590-620nm orange
575-590nm yellow 620-700 red
9
Spectral reflectance curves
See objects = light reflected from them
Reflectance curve
Achromatic colour: equal reflectance across wavelengths
White, black, grey
Chromatic colour: selective reflectance across wavelengths
10
Not all light the same
Different light sources have differing spectral
composition
Sunlight: White
light bulb: Yellow/red
11
Additive and subtractive
mixing
Lights mix additively
more wavelengths = closer to white (like sunlight)
Pigments mix subtractively
more wavelengths = closer to black
Subtractive
Additive
B absorbs Y
Y absorbs B
B & Y commonly reflect green
12
How many colours can we
perceive?
~2,000,000= 200 hues x 500 brightness levels x
20 saturations levels
Hue
Brightness
wavelength
amplitude of wave = intensity
# of photons
Saturation
Degree of white
RED vs PINK
13
Trichromatic theory of colour
perception
2 theories from the 1800s based on
psychophysical data
Trichromatic theory of colour vision
Young and von Helmholtz
Colour-matching experiments
Mix 3 pure lights (420, 560, 640) until
matches another light (500nm)
Conclusions: able to duplicate colour by
adjusting proportion
14
Trichromatic theory of colour
perception
Trichromatic theory (cont’d)
Colour vision depends on 3 receptor
mechanisms with different spectral
sensitivities
Particular wavelength stimulates 3
mechanisms to different degrees and pattern
of activity in 3 mech = perception of colour
15
Trichromatic theory: Physiology
Physiology – a century later…
3 cone visual pigments with different
absorption:
Short: 419nm
Middle 531nm
Long: 558nm
16
Colour: Its all in the ratio
Perception of colour depends upon ratio of excitation across
receptors
17
Metamers
Lights that are physically different can look
identical
How so?
Ratio of excitation across receptors is =
Same colour despite different wavelengths
Explains colour-matching experiment
Although both lights have different wavelengths, they
perceptually look the same
Metamers look the same because generate same
activation responses in 3 types of cone receptors
18
Principle of univariance
Do we need 3 receptors?
What about 1?
NO, not possible due to
principle of univariance
Varying intensity (# of
photons) can allow to have
same # of isomerized
molecules of pigments
This is why we don’t see
colour in dim light,
because rely on one ROD
pigment
What about 2?
YES but fewer colours
(see text)
More confusion btwn
colours
19
Opponent process theory of
colour
Ewald Hering
Opposing responses generated by blue and
yellow and by red and green.
Phenomenological observations
Afterimages
Simultaneous color contrast
Can’t picture reddish-green or bluish-yellow
Colour-blind: red+green; blue-yellow
20
Afterimages
21
Afterimages and
simultaneous colour contrast
Colour opposites
22
Opponent process:
Colour appearance
Rating of colour
experience for
different
wavelengths
Little co-occurrence
Reddish-green
Bluish-yellow
23
Opponent-theory
3 mechanisms: respond in opposite ways to
intensity and wavelength
Black (-) | white (+)
Red (+) | Green (-)
Blue (-) | Yellow (+)
24
Physiology: Opponent neurons
in retina and LGN
Signals from cones are transformed
early.
M retinal ganglion cells are
achromatic • dark - light
P retinal ganglion: centre / surround
are sensitive to different wavelengths
of light
• red – green • blue - yellow
25
Architecture of opponent
cells
Dual process theory
L+M–
S+ A- (sum M&L)
26
Colour and lightness
constancy
Pure wavelength information insufficient to
explain colour perception
Luminance insufficient
to explain lightness
27
Wavelengths and colour
perception
V1
Selective for the wavelength of light
However, precise wavelength of light often
bears little relationship to the perceived colour
V4
Neurons behave as if they are responding to
colours as seen by human observers
28
10 minute break
29
Depth
Of feeling? Knowledge?
Space
3D world —>2D projection on retina—>
3D perception
Need to “reconstruct” 3D world
30
Flatland:
A romance of many
dimensions
Edwin Abbott (1884)
A point, a line, a cube
31
How do we reconstruct
depth?
3 sources of information
Extraretinal oculomotor cues
Physiological/muscular feedback
Monocular cues
Pictorial
Can be recovered from one eye
Lots of them
Binocular
Disparity
2 eyes, 2 views of the world
32
Oculomotor
Afferent feedback from body
Vergence
“Convergence”
Degree of crossing as eyes
fixate
Near vs far
Accomodation
Stretching of lens to focus light
on retina
33
Monocular depth
Are 2 eyes better than
1?
Yes
Are 3 eyes better than
2?
Not many one eyed or
three eyed creatures
Nonetheless, can see
depth with 1 eye
34
Monocular cues:
Linear perspective
Parallel lines converge with distance
Converge at vanishing point (horizon)
35
Monocular cues:
familiarity and relative size
2 objects are of equal
size (familiarity)
Smaller retinal
projection—>further
away
World: Same size
Retina: Different size
36
Monocular cues:
Relative height and shadows
37
Monocular cues:
Occlusion
Layers of depth stretching out to horizon
38
Monocular cues: Atmospheric
blur and depth of focus
Blurriness
Haze
Depth of focus
In front and
behind of
fixation
39
Monocular cues:
Combine to form depth
Occlusion, relative height, and shadows
Impossible:
Conflicting cues
40
Monocular cues:
Dynamics cues
Motion parallax
Velocity = distance/time
Km/hour
As observer moves
Objects closer move faster
Greater distance across retina
Objects further move slower
E.g. looking out a train window
41
Binocular cues: Stereopsis
Why have two eyes?
Shared field of view (FOV)
Not just more = better
2 overlapping but distinct visions of the world
Sacrifice: 360 degree FOV
Gain: depth through horizontal disparity
Predators (overlap) vs prey (larger FOV)
No overlap
Substantial overlap
42
Stare at your thumb
One eye at a time
Horizontal disparity
Thumb moves side by
side
2 very different
perspectives on world
Vertical disparity?
43
Horopter:
An isodepth sphere
Horopter
Uncrossed disparity
Fixate on an object
An imaginary sphere that
defines corresponding points
on the retinas
Zero disparity
Nasal of fovea
Further in depth
Crossed disparity
Temporal of fovea
Closer in depth
Uncrossed
Fixation/
zero disparity
Retinas
Crossed
44
Remember? LGN retinal layers
Organization of LGN: Retinotopy
6 representations of retina in register
45
How do we know steropsis
produces depth perception?
Depth perception may depend solely on
“knowledge”
Monocular cues
Occlusion, familiarity etc.
Retinal disparity vs knowledge
Depth without awareness of form?
46
Random dot stereograms
Stereoscope
L & R eye shown
separate images
Random dots with
invisible disparities
Disparity alone can
result in depth
Stereoscope
Wheatstone
Crossed
47
Magic eye: Autostereograms
48
3D movies: Anaglyphs
Color filters project
different images to each
eye
49
Disparity representations in
the brain
Can’t happen at the ganglion cell layer
V1 ocular dominance columns
V1 has neurons tuned to retinal disparities
50
Part 2: Perceiving Size
Not as simple as size of stimulus on retina
Visual angle: retinal projection depends on
distance
Different physical
size
Same retinal
Projection
Bigger stimulus further away
= closer smaller stimulus
51
Size constancy
Perception of size remains constant
Despite different visual angle/retinal size
52
Size distance scaling
Perceived size = retinal image size X
distance from object
Without depth information
Perceived size = retinal image size
53
Emmert’s law
Perceived size of an after image depends
on depth perception
54
Size-depth illusions
Moon appears larger on the horizon than the sky
Same retinal size
Difference in magnitude
estimation
Horizon provides depth cues
Sky does not
Appear flattened
55
The end
56