Color, Photoreceptors Thurs. Jan. 18, 2018 1 Weeks 1 & 2 - - PowerPoint PPT Presentation

β–Ά
color photoreceptors
SMART_READER_LITE
LIVE PREVIEW

Color, Photoreceptors Thurs. Jan. 18, 2018 1 Weeks 1 & 2 - - PowerPoint PPT Presentation

COMP 546 Lecture 3 Color, Photoreceptors Thurs. Jan. 18, 2018 1 Weeks 1 & 2 image formation and measurement Weeks 3 & 4 computational modelling (early vision) Assignment 1 to be posted by end of week 3 What is light ? 3


slide-1
SLIDE 1

1

COMP 546

Lecture 3

Color, Photoreceptors

  • Thurs. Jan. 18, 2018
slide-2
SLIDE 2

Weeks 1 & 2

  • image formation and measurement

Weeks 3 & 4

  • computational modelling (early vision)
  • Assignment 1 to be posted by end of week 3
slide-3
SLIDE 3

What is light ?

3

slide-4
SLIDE 4

4

Light consists of electromagnetic waves from 400-700 nm.

slide-5
SLIDE 5

Radiance (intensity of a light ray)

5

Depends on 3D position XYZ and orientation and wavelength l

(π‘Œ, 𝑍, π‘Ž) 𝑀

𝑀

slide-6
SLIDE 6

Light Spectrum

6

For a fixed position XYZ and direction , the radiance is a function

  • nly of wavelength l.

𝑀

l

400 nm 700 nm

slide-7
SLIDE 7

Three types of spectra

7

Emission Surface Reflectance (fraction) Absorption by photoreceptors (fraction)

slide-8
SLIDE 8

Emission spectra

8

slide-9
SLIDE 9

Reflectance spectra

9

Six samples (and model fit) from standard paper color Chart. I’m not sure which six squares these plots correspond to.

For each wavelength, what is the fraction

  • f light reflected from each surface ?

https://foundationsofvision.stanford.edu/chapter-9-color

slide-10
SLIDE 10

Absorption spectra

(coming soon)

10

photoreceptors retina

slide-11
SLIDE 11

Photoreceptors: Rods and Cones

11

Cone Rod

slide-12
SLIDE 12

12

Photoreceptor response:

transduction

Light is absorbed by a pigment in photoreceptor cell. This leads to opening and closing of ion channels, causing:

  • changes in electrical potential across cell
  • membrane. (We can measure this.)
  • release of neurotransmitters which binds

to neighbor cell, i.e. communication. (We cannot measure this.)

slide-13
SLIDE 13

13

Rods

  • night (dark)
  • β€œgrey level”
  • peripheral vision
  • high noise
  • low resolution

Cones

  • day (bright)
  • color
  • central vision
  • low noise
  • high resolution (in center)
slide-14
SLIDE 14

Rods and cones: light levels

14

slide-15
SLIDE 15

Spectral sensitivity (absorption) of rods

15

Fraction of light of wavelength l absorbed by rod pigment. (Normalized to 1 for illustration purposes)

slide-16
SLIDE 16

Spectral sensitivity (absorption) of cones

16

Fraction of light of wavelength l absorbed by cone pigments. (Normalized to 1 for illustration purposes,.) L - long M - medium S - short

slide-17
SLIDE 17

β€œPrinciple of Univariance” (1D)

17

Once a photon of some wavelength is absorbed (β€œcaught”) by the photoreceptor, detailed information about spectrum is lost. How to express this mathematically?

slide-18
SLIDE 18

18

𝐹(𝑦, πœ‡)

  • spectrum of light arriving at cone 𝑦

𝐷𝑀𝑁𝑇 πœ‡

  • spectral sensitivity of a photoreceptor

(either L, M, or S)

𝐽𝑀𝑁𝑇 𝑦 = 𝐷𝑀𝑁𝑇 πœ‡ 𝐹 𝑦, πœ‡ π‘’πœ‡

πœ‡ 𝐷𝑀 πœ‡ πœ‡ 𝐷𝑁 πœ‡ πœ‡ 𝐷𝑇 πœ‡

slide-19
SLIDE 19

19

=

𝐷𝑀𝑁𝑇 πœ‡ 𝐹 𝑦, πœ‡ 𝐽𝑀𝑁𝑇 𝑦

3 x 1 3 x N N x 1

ASIDE: The photoreceptor β€œresponse” is a function of 𝐽𝑀𝑁𝑇 𝑦 .

slide-20
SLIDE 20

Metamers

20

Two different spectra can map to the same LMS triplet, i.e. Such spectra are visually indistinguishable, and are called β€œmetamers”.

𝐷𝑀𝑁𝑇 𝐹1 = 𝐷𝑀𝑁𝑇 𝐹2

slide-21
SLIDE 21

Color blindness

21

=

𝐷𝑀𝑁𝑇 πœ‡ 𝐹 𝑦, πœ‡ 𝐽𝑀𝑁𝑇 𝑦

Three types of color blindness depend on which

  • f three cone photopigments are missing.
slide-22
SLIDE 22

Cone mosaic

22

Normal Color blind (e.g. missing L cones)

slide-23
SLIDE 23

Rod (night) vision is an extreme case of metamerism

23

slide-24
SLIDE 24

Rod sensitivity

24

=

𝐷𝑠𝑝𝑒 πœ‡ 𝐹 𝑦, πœ‡ 𝐽𝑠𝑝𝑒 𝑦

slide-25
SLIDE 25

Application: RGB images and color displays

25

(1, 1, 1) (1, 1, 0) (1, 0, 1) (0, 0, 1) (0, 1, 1) (0, 1, 0) (0, 0, 0)

slide-26
SLIDE 26

Emitted spectra from color displays

26

=

𝑄𝑆𝐻𝐢

RGB

in [0,1]

𝐹𝑆𝐻𝐢 𝑦

N x 1 N x 3 3 x 1 When properly drawn, peaks correspond to P matrix on right.

computer monitor or projector

slide-27
SLIDE 27

27

=

𝐷𝑀𝑁𝑇 πœ‡ 𝐽𝑀𝑁𝑇 𝑦

3 x 1 3 x N N x 3 3 x 1

𝑄𝑆𝐻𝐢

RGB

in [0,1]

Monitor, projector spectra

3 x 3 transform from RBG pixel values to cone absorptions cone absorption emission

slide-28
SLIDE 28

Anaglyph 3D Displays

28

Anaglyph (definition): a photograph with the two images superimposed and printed in different colors, producing a 3D effect when viewed through correspondingly colored filters. transmits red (long) transmits cyan (short)

slide-29
SLIDE 29

left and right eye image

29

white (1,1,1) cyan (0,1,1) red (1,0,0) black (0,0,0)

slide-30
SLIDE 30

30

Emission from 3DTV Absorption by photoreceptors (fraction) Transmission (fraction)

slide-31
SLIDE 31

Simplified anaglyph model

31

emitted * transmitted (RBG) (filter) Examples: (1, 0, 0) = (1, 0, 0) * (1, 0, 0) (1, 0, 0) = (1, 1, 1) * (1, 0, 0) (0, 0, 0) = (0, 1, 1) * (1, 0, 0) (0, 0, 0) = (1, 0, 0) * (0, 1, 1) (0, 1, 1) = (0, 1, 1) * (0, 1, 1) (0, 1, 1) = (1, 1, 1) * (0, 1, 1)

slide-32
SLIDE 32

Simplified anaglyph model

32

emitted * transmitted (RBG) (filter) Examples: (1, 0, 0) = (1, 0, 0) * (1, 0, 0) (1, 0, 0) = (1, 1, 1) * (1, 0, 0) (0, 0, 0) = (0, 1, 1) * (1, 0, 0) (0, 0, 0) = (1, 0, 0) * (0, 1, 1) (0, 1, 1) = (0, 1, 1) * (0, 1, 1) (0, 1, 1) = (1, 1, 1) * (0, 1, 1)

slide-33
SLIDE 33

Summary of today

  • Types of spectra
  • Rods and cones
  • Metamers and color blindness
  • Color displays
  • Temporal issues
  • Spatial issues (next class)

33

slide-34
SLIDE 34

34

Photoreceptor response to a brief flash of light

(depends on intensity, duration, background level)

0 100 time (ms) flash of light

  • 40 mV
  • 50 mV

Response

slide-35
SLIDE 35

Response of photoreceptor*

(non-linear)

35

100 Response (%) log brightness of light

saturation *You get the same sigmoidal behavior for cameras.

slide-36
SLIDE 36

Response of photoreceptor*

(non-linear)

36

100 Response (%) log brightness of light

Different curves are for different background intensity. saturation *You get the same sigmoidal behavior for cameras.

slide-37
SLIDE 37

Adaptation

37

Look at dot for 30 sec Then, look at this dot.

slide-38
SLIDE 38

Adaptation Time Scales

  • fraction of a second

as we scan a scene with eye movements

  • minutes

as we change environments light adaptation: rods to cones dark adaptation: cones to rods

38

slide-39
SLIDE 39

Pupil Response

39

Partly compensates for changes in average light level over the whole image.

Diameter of pupil ~2 - 8 mm.

This is only a small contribution to huge operating range of the visual system.