Vision IV. Overview of Topics. Evolution of Vision. Overview of Topics. Colour Vision

Similar documents
Vision IV. Overview of Topics. Overview of Topics. Colour Vision

Slide 1. Slide 2. Slide 3. Light and Colour. Sir Isaac Newton The Founder of Colour Science

excite the cones in the same way.

Question From Last Class

Additive. Subtractive

Spectral colors. What is colour? 11/23/17. Colour Vision 1 - receptoral. Colour Vision I: The receptoral basis of colour vision

Sensation, Part 4 Gleitman et al. (2011), Chapter 4

Light and Colour. Light as part of the EM spectrum. Light as part of the EM spectrum

What is Color. Color is a fundamental attribute of human visual perception.

COLOR and the human response to light

Color Perception. Color, What is It Good For? G Perception October 5, 2009 Maloney. perceptual organization. perceptual organization

COLOR. and the human response to light

Image and video processing (EBU723U) Colour Images. Dr. Yi-Zhe Song

CS 565 Computer Vision. Nazar Khan PUCIT Lecture 4: Colour

Colour. Why/How do we perceive colours? Electromagnetic Spectrum (1: visible is very small part 2: not all colours are present in the rainbow!

Color and Color Model. Chap. 12 Intro. to Computer Graphics, Spring 2009, Y. G. Shin

Color Science. What light is. Measuring light. CS 4620 Lecture 15. Salient property is the spectral power distribution (SPD)

Colour. Cunliffe & Elliott, Chapter 8 Chapman & Chapman, Digital Multimedia, Chapter 5. Autumn 2016 University of Stirling

Colour. Electromagnetic Spectrum (1: visible is very small part 2: not all colours are present in the rainbow!) Colour Lecture!

Reading. Foley, Computer graphics, Chapter 13. Optional. Color. Brian Wandell. Foundations of Vision. Sinauer Associates, Sunderland, MA 1995.

Victor Ostromoukhov Université de Montréal. Victor Ostromoukhov - Université de Montréal

Test 1: Example #2. Paul Avery PHY 3400 Feb. 15, Note: * indicates the correct answer.

Figure 1: Energy Distributions for light

iris pupil cornea ciliary muscles accommodation Retina Fovea blind spot

Color Perception. This lecture is (mostly) thanks to Penny Rheingans at the University of Maryland, Baltimore County

Psy 280 Fall 2000: Color Vision (Part 1) Oct 23, Announcements

Color & Graphics. Color & Vision. The complete display system is: We'll talk about: Model Frame Buffer Screen Eye Brain

Future Electronics EZ-Color Seminar. Autumn Colour Technology

To discuss. Color Science Color Models in image. Computer Graphics 2

Color Science. CS 4620 Lecture 15

CMPSCI 670: Computer Vision! Color. University of Massachusetts, Amherst September 15, 2014 Instructor: Subhransu Maji

Color. Bilkent University. CS554 Computer Vision Pinar Duygulu

Computer Graphics Si Lu Fall /27/2016

University of British Columbia CPSC 314 Computer Graphics Jan-Apr Tamara Munzner. Color.

Interactive Computer Graphics

LECTURE 07 COLORS IN IMAGES & VIDEO

Colors in Images & Video

Reading for Color. Vision/Color. RGB Color. Vision/Color. University of British Columbia CPSC 314 Computer Graphics Jan-Apr 2013.

Color. Fredo Durand Many slides by Victor Ostromoukhov. Color Vision 1

any kind, you have two receptive fields, one the small center region, the other the surround region.

Today. Color. Color and light. Color and light. Electromagnetic spectrum 2/7/2011. CS376 Lecture 6: Color 1. What is color?

skip chap. 8 for now Chap. 9 Color (continued) Lecture 19 Tuesday, October 26

Color. Color. Colorfull world IFT3350. Victor Ostromoukhov Université de Montréal. Victor Ostromoukhov - Université de Montréal

PERCEPTUALLY-ADAPTIVE COLOR ENHANCEMENT OF STILL IMAGES FOR INDIVIDUALS WITH DICHROMACY. Alexander Wong and William Bishop

CS 544 Human Abilities

IFT3355: Infographie Couleur. Victor Ostromoukhov, Pierre Poulin Dép. I.R.O. Université de Montréal

Digital Image Processing. Lecture # 8 Color Processing

PERCEIVING COLOR. Functions of Color Vision

Vision. PSYCHOLOGY (8th Edition, in Modules) David Myers. Module 13. Vision. Vision

Lecture 6 6 Color, Waves, and Dispersion Reading Assignment: Read Kipnis Chapter 7 Colors, Section I, II, III 6.1 Overview and History

Color Deficiency ( Color Blindness )

CS6640 Computational Photography. 6. Color science for digital photography Steve Marschner

OPTO 5320 VISION SCIENCE I

Frequencies and Color

For a long time I limited myself to one color as a form of discipline. Pablo Picasso. Color Image Processing

color basics theory & application Fall 2013 Ahmed Ansari Communication Design Fundamentals

Capturing Light in man and machine

Color and perception Christian Miller CS Fall 2011

Color images C1 C2 C3

The Principles of Chromatics

Capturing Light in man and machine

Fig Color spectrum seen by passing white light through a prism.

Understand brightness, intensity, eye characteristics, and gamma correction, halftone technology, Understand general usage of color

Creative Computing II

Colour, Vision & Perception

Introduction. The Spectral Basis for Color

Comparing Sound and Light. Light and Color. More complicated light. Seeing colors. Rods and cones

Lecture Color Image Processing. by Shahid Farid

Color, Vision, & Perception. Outline

Capturing Light in man and machine

Color Image Processing. Gonzales & Woods: Chapter 6

LECTURE III: COLOR IN IMAGE & VIDEO DR. OUIEM BCHIR

Reading. Lenses, cont d. Lenses. Vision and color. d d f. Good resources: Glassner, Principles of Digital Image Synthesis, pp

Multimedia Systems Color Space Mahdi Amiri March 2012 Sharif University of Technology

Digital Image Processing COSC 6380/4393. Lecture 20 Oct 25 th, 2018 Pranav Mantini

19. Vision and color

SYDE 575: Introduction to Image Processing. Adaptive Color Enhancement for Color vision Deficiencies

Wireless Communication

10.2 Color and Vision

Mahdi Amiri. March Sharif University of Technology

Capturing Light in man and machine. Some figures from Steve Seitz, Steve Palmer, Paul Debevec, and Gonzalez et al.

Visual Perception. human perception display devices. CS Visual Perception

Multimedia Systems and Technologies

Color Theory. Additive Color

Conceptual Physics Fundamentals

III: Vision. Objectives:

Additive Color Synthesis

Reflectance curves of some common foods. Spectral colors. What is colour? 11/4/11

Color. April 16 th, Yong Jae Lee UC Davis

Digital Image Processing Color Models &Processing

Light. intensity wavelength. Light is electromagnetic waves Laser is light that contains only a narrow spectrum of frequencies

Capturing Light in man and machine

The Physiology of the Senses Lecture 1 - The Eye

check it out online at

This question addresses OPTICAL factors in image formation, not issues involving retinal or other brain structures.

Color Image Processing

LIGHT AND LIGHTING FUNDAMENTALS. Prepared by Engr. John Paul Timola

Color April 16 th, 2015

Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester

Part I: Color Foundations The Basic Principles of COLOUR theory

Transcription:

Overview of Topics Vision IV Colour Vision Chapter 11 in Chaudhuri "Avoid vertebrates because they are too complicated, avoid colour vision because it is much too complicated, and avoid the combination because it is impossible."!!!!advice given to his students by!!!!nobel-prize winner H. Keffer Hartline 1 2 Overview of Topics Evolution of colour vision Physics of Colour Biological Foundations of Colour Vision Perceptual Aspects of Colour Vision Evolution of Vision Eyes evolved 500 Myr ago in the Cambrian Explosion Figure shows a sequence of increasingly complex eyes All exist in creatures today and are thought to be the precursors of advanced chambered eyes like ours 3 4

Evolution of Multiple Cone Types Early fish ( 360 Myr ago) already had 4 cone types! Our nocturnal rat-like mammal ancestors had just 2, as do many modern mammals Some primates, including humans, have 3 Many birds and fish have 4 or more, probably aiding them mainly in mate selection Functions of Colour Vision Aids in segmenting the scene into objects This makes it easier to find food/prey and avoid predators via camouflage breaking Helps to classify objects, often along subtle dimensions (ripe/unripe, healthy/unhealthy) Aids in feeding and mate selection 5 6 Colour Vision Helps Segmentation Crouching Tiger 7 8

Hidden Dragon Plants Take Advantage of Colour Vision 9 10 Classification Frugivorous monkeys use colour vision to pick out ripe fruit from green foliage Classification Colour is important in mate selection, especially amongst birds and fish, where it is an indicator of male health. Hot Not Folivorous monkeys use colour vision to pick out new red leaves from old green ones 11 Among Zebra Finches, the ladies like a nice bright orange cheek patch 12

Evolution of Colour Vision Questions Due to different evolutionary pressures, different species have radically different colour vision capacities (and radically different sensory capacities in general) What are some of the functions of colour vision? How would you explain red to a cat? 13 14 A Little Light Philosophy Colour Science Is the colour in the object, or in the observer? A classic bad question. It is in both. Colour experiences result from an interaction between physical properties of objects and light, and the physiology of an observing organism 15 16

Light (reprise) A form of electromagnetic (EM) radiation (along with gamma rays, UV light, radio, etc.) EM radiation varies by: Wavelength Intensity Polarity Wavelength Abbreviated λ (lambda) Measured in units of distance, such as angstroms or nanometers (nm, 1x10-9 ) Visible light (to humans) ranges from 400 nm to 700 nm Variations in wavelength give rise to colour experience, but relationship is complex. 17 18 Wavelength Laser Lights Long λ (perceived as red) Medium λ (perceived as green) Short λ (perceived as blue) A laser emits a single wavelength, so is a nice simple model to look at first (analogous to a pure tone in hearing) Its colour appearance will vary across the spectrum with wavelength 19 20

Example Emission Spectra Colour Circle The colour circle shows the colour experience produced by each individual λ These are called spectral colours, as they can be produced by a single λ Purple is a special case, an example of a nonspectral colour. No single λ can produce a nonspectral colour experience 21 22 Nonspectral Colours Questions Spectral colours can be produced by a single λ Non-spectral colours can only be produced by a mixture of λs. Examples: Purple: Mix 400 nm and 700 nm lights Pink: Mix 460 nm and 600 nm (white) with a bit of 700 nm (red) How many λs in a laser light? What about all other lights? What is a spectral colour? What about a nonspectral colour? 23 24

Additive Colour Mixes A colour circle can also be used to show the appearance of a mix of λs Draw a line between the 2 λs being mixed If the λs are of equal intensity, the midpoint shows the appearance Otherwise, pick a point along the line at a proportional distance from the two components Complementary Colours Any 2 λs across the centre from one another are complementary colours Mixing them in equal proportions produces an achromatic colour experience (i.e., black-grey-white) Achromatic colours are examples of nonspectral colours 25 26 Full Spectrum White Sunlight and other full spectrum white light sources produce equal numbers of photons across the λs spectrum Thus they produce all the spectral colours mixed together One can think of this as all of the pairs of complementary colours mixed together, such that an achromatic colour experience occurs Other Ways to Make White Fluorescents look white because they combine a few different λs that together are complementary (refer back to the emission spectrum) Any two sets of λs that produce the same colour experience are said to be metamers Thus, all pairs of complementary λs are examples of metamers, because they all produce white 27 28

Metamers Any two lights with different λ contents that produce the same colour experience are called metamers Example: Yellow can be produced by either 1) 600 nm light alone 2) 550 nm (green) + 650 nm (red) Monitors & The RGB Colour Space Monitors produce colours by varying the intensity of three phosphors in each pixel The phosphors are red, green, and blue, abbreviated RGB Each one typically has 256 levels of intensity (0-255) 29 30 Monitors & the RGB Colour Space Not One λ, But Many Each phosphor type emits a range of λs, no just one (they re not lasers) the colours they individually produce are in the body of the circle, not the edge The colours they can mix to produce are limited to those inside the RGB triangle (e.g., no laser green, and few orange shades) R=255 G=128 B=128 R=255 G=000 B=000 R=255 G=255 B=255 Natural lights have a range of λs, each with its own intensity. Only laser lights approximate a single λ A light s subjective colour is determined (in part) by its intensity spectrum 31 32

Example Intensity Spectra Example Emission Spectra Intensity 33 34 Intensity Spectra of RGB Phosphors Additive Primary Colours Red, Green and Blue are known as the additive primaries Nearly any colour can be made by adding them in the right proportions 35 36

Questions What s an emission spectrum? What s an intensity spectrum? What are metamers? Colour & Objects To this point, we ve mostly discussed emitted lights But most light we see is reflected from objects As we ve discussed previously, there are two main types of reflection: Specular & Body 37 38 Body Reflection Specular Reflection Occurs when photons enter an object, and are dispersed by its molecules Light is bounced off in all directions as hemispheric wavefronts This is primarily responsible for the colour appearance of objects Occurs when photons bounce off the surface of an object Light is bounced off at angle opposite the angle of incidence This is primarily responsible for the degree of glossiness of an object 39 40

Reflectance Spectra Reflectance Spectra To precisely describe an object s reflection properties, we use a reflectance spectrum This shows the % of photons of each wavelength that are reflected To know what colour the object will appear to have, we also need to know the emission spectrum of the light source shining on it Will a tomato always look red? 41 42 Emission, Reflectance, & Colour Signal The photons that come off an object are the product of the object s reflectance spectrum and the illuminant s emission spectrum This, along with the physiology of the visual system, determines the colour appearance Subtractive Colour Mixing Rules of colour mixture are different when mixing coloured objects (e.g., dyes, inks, paints) This is because objects absorb photons rather than emitting Sunlight 43 44

Subtractive Colour Primaries Subtractive Colour Mixing The subtractive colour primaries are cyan, magenta and yellow Note how these are the secondary colours from additive colour mixing, and vice versa Just as with additive colour mixing, one can add the subtractive primaries to produce a range of colours. However, it can be hard in practice to produce a deep black colour Therefore, printers tend to add black* and work in CMYK colour space * Note that black is denoted by K because B is taken (for blue in RGB colour space) 45 46 Questions What s the relationship between emission spectrum, reflectance spectrum, and colour spectrum? If you mix all the paints in your paint set, what colour do you get? Why? A Perceptual Colour Space: HSV RGB and CYMK represent the physics of colour. A more psychologically-based colour space is HSV: Hue (red, orange, violet): What we normally think of when we say colour Saturation (red vs. pink): Purity, vividness High = vivid. Low = washed out Value (bright red vs. dark red): aka brightness or lightness or intensity 47 48

Relating the Physical & Perceptual Hue, saturation and value each relate to a characteristic of a light s energy spectrum (or colour signal ) Hue: Position along spectrum Saturation: Narrowness of spectrum Value: Total area of spectrum Hue Spectral Position Moving the colour signal up/down the λ spectrum changes the perceived hue 49 50 Saturation Spectral Width Making the colour signal narrower increases saturation (vivid), making it wider decreases saturation (washed out) Value Spectral Area Making the colour signal larger in area makes the colour brighter/lighter.3.2. 1 Hue 2. 51 52

The HSV Colour Space Questions The difference between blue and red is a difference in, while the difference between red and pink is a difference in. A laser light exhibits maximum. 53 54 Theories of Colour Vision Theories of Colour Vision Thesis: Trichromacy, there are three colour mechanisms. Antithesis: Colour-opponency; No, there are four colour mechanisms in opposing pairs Synthesis: Modern colour vision models show that both trichromacy and opponency are correct at different stages. 55 56

Trichromatic Theory of Colour Vision Proposed by Young and Helmholtz (1800s) Suggested that three different receptor mechanisms are responsible for colour vision Behavioral evidence: Colour-matching experiments show that observers need three wavelengths to match a comparison field to a test field 57 Colour-matching Experiment 525 nm Test field 600 nm Comparison field For dichromats with only two cones, only two wavelengths are needed to match the test field (these two greens look the same to them). 59 550 nm Colour-matching Experiment 525 nm Test field 600 nm 400 nm 550 nm Comparison field Observer adjusts the intensity of the three colours that combine to make the comparison field, trying to match the test field. It turns out this can be done with any three wavelengths in the comparison field 58 Physiological Evidence for Trichromatic Theory Researchers measured absorption spectra of visual pigments in receptors (1960s) They found pigments that responded maximally to: Short wavelengths (440nm) Medium wavelengths (530nm) Long wavelengths (560nm) Later researchers found genetic differences for coding proteins for the three pigments (1980s) 60

Opponent-Process Theory of Colour Vision Proposed by Hering (1800s) Said that colour vision is caused by opposing responses generated by blue/yellow, green/red, and white/black. Behavioural evidence: Colour afterimages and simultaneous colour contrast show the opposing pairings Types of colour blindness are red/green & blue/yellow 61 62 Table 7.3 Results of afterimage and simultaneous contrast demonstration Opponent-Process Theory of Colour Vision Opponent-process mechanism proposed by Hering Three mechanisms - red/green, blue/yellow, and white/black The pairs respond in an opposing fashion, such as positive to red and negatively to green (correct, except for white/black) These responses were believed to be the result of chemical reactions in the retina 63 64

Physiology of Opponent-Process Figure 7.19 The three opponent mechanisms proposed by Hering. Researchers performing single-cell recordings found colouropponent neurones in the retina, LGN and V1. These respond in an excitatory manner to one of four colours an inhibitory manner another of the four colours. Organized along red-green and yellow-blue axes (approximately). There are also achromatic (no colour preference) neurones in these areas. 65 66 Synthesis Responses of opponent cells in the monkey s lateral geniculate nucleus. These cells respond in opposite ways to blue and yellow (B+Y- or Y+B-) and to red and green (G+R- or R+G-). Both theories are correct. Each describes physiological mechanisms in the visual system Trichromatic theory explains the responses of the cones in the retina Opponent-process theory explains responses of colour-opponent neurones in retina, LGN and V1. 67 68

Questions What are some sources of evidence for trichromatic theory? What is some evidence for opponent process theory? Which one is correct? How Do Cones Provide Information About λ? Does each cone somehow transmit info about the λ of the photons it absorbs? No. A photoreceptor responds the same way when it absorbs a photon, regardless of that photon s λ This is known as the principle of univariance Thus, a receptor can only signal how many photons it has absorbed, not their λs 69 70 How Do Cones Provide Information About λ? For any given colour signal, the cones types will absorb different #s of photons Because of univariance, we must compare across cone types to get information about λ It is the relative differences in the firing rates of cones that provide information about λ That is, this is a cross-fibre coding scheme How Do Cones Provide Information About λ? Colour Signal Colour Signal 71 72

How Do Cones Provide Information About λ? Physiological Basis of Metamers Cones don t provide perfect information about λ, or there would be no metamers Metamers occur because many patterns of λs can produce the same pattern of activation across cone types Some of trichromatic theory s strongest evidence was its ability to explain metamers. 73 74 Physiological Basis of Metamers Questions How do cones provide information about λ? Why do metamers happen? 75 76

Why 3 Cone Types? Monochromacy: 1 cone type provides no colour vision due to univariance Dichromacy: 2 cone types provides crude colour vision (red/green OR blue/yellow) Trichromacy: 3 cones types provides better colour vision (red/green AND blue/yellow) More cones provides better colour discrimination, but at the expense of acuity Why Not More Cone Types? More cones would provide better colour discrimination Might allow frequency analysis, as with sound But each receptor type would have low density, thus low acuity 77 78 Why Not 16 Cone Types? 1 Receptor Type Monochromacy: True colour blindness. All λs look the same hue. One receptor type cannot provide any information about λ independent of intensity Advantage is high acuity and/or sensitivity (e.g. rod system in humans) "The eye has no sense of harmony in the same meaning as the ear. There is no music to the eye." -Helmholtz 79 80

2 Receptor Types Dichromacy: Provides a single dimension of colour vision. Information about λ can be obtained by comparing S cone to L cone. Long λs will activate L cone more, while short λs will activate S cone more. With one photoreceptor type, two wavelengths of light might produce different outputs... 81 82...or they might produce the same outputs. It depends on the intensity (number of photons) at each λ. Thus, a one-cone system can t differentiate between intensity and wavelength differences. 83 84

3 Receptor Types Allows multiple dimensions of colour sensitivity In humans, red/green and blue/yellow Questions Why can one cone type not provide colour vision? What are the two basic dimensions of human colour vision? 85 86 Colour Deficiency Most humans are normal trichromats with S, M, and L cones having typical absorption spectra But other forms of colour vision exist in humans, including: Rod Monochromacy Cone Monochromacy Dichromacy Anomalous trichromacy Unilateral dichromacy. L, M, & S subtypes (aka Protan, Deutan and Tritan) Rod Monochromacy Extremely rare ( 1:33000) Only rods and no functioning cones Perceive only in white, grey & black tones True colour-blindness Poor visual acuity ( 20/150) Very sensitive to bright light ( day blind ) 87 88

Cone Monochromacy Extremely rare (incidence unknown) One type of cone (plus rods). 3 varieties: S, M, or L-cone monochromacy. Individuals have generally normal vision but can discriminate no hues. Three types: Dichromacy Protanopia: Missing L cones Deuteranopia: Missing M cones Tritanopia: Missing S cones 89 90 Dichromacy: Protanopia Missing L cone 1% of males and.02% of females! Individuals see short-wavelengths as blue Neutral point occurs at 492nm ( halfway between M and S cone peaks) Above neutral point, they see yellow Dichromacy: Deuteranopia Missing M cone 1% of males and.01% of females Individuals see short-wavelengths as blue Neutral point occurs at 498nm ( halfway between peaks of S and L cones) Above neutral point, they see yellow 91 92

Dichromacy: Tritanopia Missing S cone.002% of males &.001% of females Individuals see short wavelengths as green Neutral point occurs at 570nm ( halfway between peaks of M and L cones) Above neutral point, they see red Note error in text 93 94 Anomalous Trichromacy 3 cone types, but pigments are not isolated (e.g., L cones may have some M cone pigment in them, or vice versa) Need 3 λs to match any colour, but their proportions are different than for normal trichromats. Poorer hue discrimination Anomalous Trichromacy Comes in three types: No neutral point. Protanomaly (L cones): 1%,.01% Deuteranomaly (M cones): 6%,.4% Tritanomaly (S cones):.01% & 95 96

Pseudoisochromatic Plates Cap-arrangement Tests Ishihara pseudoisochromatic plate One method for testing for colour deficiency. Why is it made of bubbles of different shades? Farnsworth-Munsell DM-15 cap arrangement test Task is to place coloured caps in a smooth progression of colours. Colour-deficient observers make characteristic patterns of mistakes 97 98 Cap-arrangement Test Questions If your friend seems to have terrible fashion sense, mixing orange pants with a green shirt, he most likely suffers from: a) Being male b) Being a hipster c) Deuteranomalia d) All of the above 99 100

Physiological Basis For Colour Opponency Trichromacy is correct at the level of the retina: We have 3 cone types But behavioural evidence shows that we have 4 colour primaries in opponent pairs: Red/Green and Blue/Yellow How can this be? Colour-Opponent Neurones Found in retina, LGN and V1, these neurones are excited by one range of λs and inhibited by another Four types exist: R+/G-, G+/R-, B+/Y-, Y+/B- How do these arise from the L, M, and S cones? 101 102 Colour-Opponent Neurones From Trichromacy to Opponency 103 104

From Trichromacy to Opponency Achromatic Cells While some RGCs are colour opponent, many are not These are the ON/OFF and OFF/ON cells we saw earlier These most likely add up L and M cone responses in both centre and surround (recall that S cones have little to do with luminance perception) 105 106 From Trichromacy to Opponency Questions What kinds of cones are found in the surround of a B/Y RGC s receptive field? What kinds of cones are found in the surround of an ON/OFF RGC receptive field? 107 108

Colour Processing in Cortex Not fully understood There are colour opponent cells in V1 and beyond There are also double-opponent cells The Isoluminance Problem Isoluminance means same luminance For instance, red and green stripes can be made to have the same luminance We can see the edge between such stripes, but how? Achromatic cells would not respond, and neither would colour opponent cells 109 110 Colour-Opponent Cells Don t Respond at Isoluminance Double-Opponent Cells In V1, cells are found that combine spatial and chromatic opponency For example, R+G-/G+R- cells are excited by red light on centre and green on surround Other configurations, exist: G+R-/R+G- B+Y-/Y+B- Y+B-/B+Y- Most are found in cytochrome oxidase blobs 111 112

Perceptual Aspects of Colour Vision Wavelength Discrimination Thresholds Difference detection Colour contrast Colour constancy 113 114 Saturation Discrimination Thresholds Colour Contrast The background of an object (among many other things) can affect colour perception 115 116

Colour Constancy Perception of colours is relatively constant despite changing light sources Sunlight has approximately equal amounts of energy at all visible wavelengths Tungsten lighting has more energy in the longwavelengths Objects reflect different wavelengths from these two sources 117 118 The Mystery of the Green Sweater Figure 7.24 The reflectance curve of a sweater (green curve) and the wavelengths reflected from the sweater when it is illuminated by daylight (white) and by tungsten light (yellow). Possible Causes of Colour Constancy Chromatic adaptation - prolonged exposure to chromatic colour leads to: Receptors adapt when the stimulus colour selectively bleaches a specific cone pigment Sensitivity to the colour decreases adaptation occurs to light sources leading to colour constancy 119 120

Observers shown sheets of coloured paper in 3 conditions: Baseline - paper and observer in white light Observer not adapted - paper illuminated by red light; observer by white Uchikawa et al. Observer adapted - paper and observer in red light 121 122 Uchikawa et al. Results showed that: Baseline - green paper is seen as green Observer not adapted - perception of green paper is shifted toward red Observer adapted - perception of green paper is slightly shifted toward red (i.e, partial colour constancy) Possible Causes of Colour Constancy Grey World assumption: Visual system may discount overall hue of scene. Evidence: Colour constancy works best when an object is surrounded by many colours Memory and colour: Past knowledge of an object s colour can have an impact on colour perception Memory for colour is poor, so we don t notice slight changes caused by illumination changes 123 124

Hue Depends on Intensity Questions What are some possible causes of colour constancy? Which of these does the Uchikawa experiment demonstrate? 125 126 THE END (of the course!) 127