Basic principles of colour perception

  •    The more light a surface absorbs the darker it appears, the colour we perceive depends on the wavelengths of the light that is reflected by the object and reaches our eyes

Three steps to colour perception

  1. Detection: wavelengths must be detected
  2. Discrimination: we must be able to tell the difference between one wavelength and another
  3. Appearance: we want to assign perceived colours to lights and surfaces in the world, and we don’t want them to change dramatically as the viewing conditions change
  4. Colour detection
    • Three types of cone photoreceptors (S-cones peak at about 420nm, M-cones peak at about 535nm, L-cones peak at about 565nm), are sensitive to light of different wavelengths. Scones are relatively rare, and they are less sensitive than M- and L-cones.
    • Cones work at photopic (daylight) light levels
    • Rods work at scotopic (dimmer) light levels
  5. Colour discrimination
    • The output of a single photoreceptor is completely ambiguous, it can respond to two different wavelengths at the same intensity with the exact same response (e.g. 450 and 625nm for M-cones) and if you adjust white light to a certain intensity it will produce the same response again
    • Problem of univariance: the fact that an infinite set of different wavelength-intensity combinations can elicit exactly the same response from a single type of photoreceptor. One photoreceptor type cannot make colour discriminations based on wavelength o In dim light only rods will respond, and there is only one kind of rod photoreceptor with one type of photopigment (rhodopsin), thus they all have the same sensitivity to wavelength making it impossible to perceive colour in dim light

The trichromatic solution

  • We can detect differences between wavelengths or mixtures of wavelengths because we have three kinds of cone types
  • Trichromatic theory of colour vision: the combination of the responses from different photoreceptors creates a unique set which we use as the basis for colour vision:

 

  • If the intensity of light increases, the responses of the individual cones will change, but the relationships between the different cones does not, and those relationships will define our response to the light and eventually the colour that we see

Metamers

  • Metamers: different mixtures of wavelengths that look identical. More generally, any pair of stimuli that are perceived as identical in spite of physical differences. The nervous system only knows what the cones tell it, if a mixture produces the same cone output as the single wavelength of yellow light than the mixture and the single wavelength must look identical, thus as yellow.

 

However:

o Colour mixture is a mental event, not a change in the physics of light o Only just the right mixture of red and green will be perceived as yellow

The history of trichromatic theory

  • To exactly match the reference colour, the mixture must be made-up of at least three different primary wavelengths, this observation led to the conclusion that the mechanism which enables humans to perceive colour must contain three different colour photoreceptor

A brief digression into lights, filters and finger paints

  • Additive colour mixture: we take on wavelength or set of wavelengths and add them together (red + green = yellow), in the perception of colour the effects of the two lights are added together o Pointillism: creating many hues by placing small spots of just a few colours in different textures, which will be visible up close, but when looked at from a distance will appear to form a uniform colour
  • Subtractive colour mixture: if pigments A and B are mixed, some of the light shining on the surface will be subtracted by A and some will be subtracted by B, only the remainder contributes to the perception of colour o Paint

o Colour filters: the filter subtracts all wavelengths except the ones which it is made-up of from the light (e.g. a yellow filter will subtract all the wavelengths except for the ones making up yellow, causing the white light to appear yellow)

From retina to brain: repackaging the info

  • Info about colour: the nervous system computes the difference between cone responses from the L- and M-cones (L-M)
  • Info about the intensity of light: the nervous system adds the response from the L- and Mcones
  • ([L+M]-S): …?

Cone-opponent cells in the retina and LGN

  • Cone-opponent cell: a cell type in the retina, LGN and visual cortex, that, in effect, subtracts/adds one type of cone input from another o (L+M), (L-M), ([L+M]-S)
  1. Colour appearance
  • Light reaching any part of the retina will be translated into three responses, one for each local population of cones
  • Colour space: the three-dimensional space, established because colour perception is based on the outputs of three cone types, that describe the set of all colours
  • Achromatic colours: arise when the amount of red is equal to the amount of green and blue light (black is 0, whit is 255, grey in between)
  • Three dimensional colour space (HSB-system):
    • Hue: the chromatic aspect of colour
    • Saturation: the chromatic strength of a hue, white has zero saturation, pink I more saturated and red is fully saturated
    • Brightness: the perceptual consequence of the physical intensity of a light
  • Nonspectral hues: colours like purplish magentas between red and blue near the end of the hue strip which cannot be found in the colour spectrum, due to the fact that they result from mixtures of wavelengths instead of single wavelengths

Opponent colours

  • Opponent colour theory: perception of colour is based on the output of three mechanisms, each of them resulting from an opponency between two colours: red-green, blue-yellow and black-white. Figure 5.14 page 128
  • Hue cancellation: start with a light that is yellowish green, by adding a blue light you can eliminate this yellowness. The right amount of blue can be measured, so that all the yellow trace can just be removed
  • Unique hues: hues that can be described with only a single colour term figure 5.16 page 130 o All the very long wavelengths look red
    • The two crossings of the red-green function provide the loci of unique blue and unique yellow
    • The point where the blue-yellow function crosses from positive to negative is the locus of unique green
  • Cone opponency
    • ([L+M]-S): adding and subtracting cone sensitivities to produce these cone-opponent cells results in cells that respond maximally along an axis extending from a purplish hue to a yellowish greenish hue
    • (L-M): L-cone end of the axis is near perceptual red, M-cone end is a bluish green, these endpoints are called cardinal directions -Figure 5.17 page 131!!!

Colour in the visual cortex

  • Single opponent cell: refers to a cone-opponent cell, this cell would have a R+ center and a G- surround, and thus this cell would be excited by a redder hue in its center and inhibited by a greener hue in its surround; conveys info about the colour of a broad area
  • Double-opponent cell: a cell type, found in the visual cortex, in which one region is excited by one cone type, combination of cones or colour and inhibited by the opponent cones or

colour (e.g. R+/G-). Another adjacent region would be inhibited by the first input and excited by the second (R-/G+); conveys info about chromatic edges o The cell described here would have a R+/G- center and a R-/G+ surround, and thus would be excited by redder hues in its center and greener hues in the surround and inhibited by greener hues in its center and redder hues in its surround –      Separate pathway for colour perception?!

  • Blobs in V1 that seem to respond to colour rather than orientation, these blobs would then sent their output to thin strip regions in V2 and from there to V4, which would have cells that respond not to wavelength but perceived colour
  • Achromatopsia: the inability to perceive colour, patients might be able to find the boundaries between regions of different colours but they cannot report what those colours might be

Adaptation and afterimages

  • Afterimages: a visual image after the stimulus has been removed
  • Adapting stimulus: a stimulus whose removal produces a change in visual perception or sensitivity
  • Negative afterimage: an afterimage whose polarity is the opposite of the original stimulus. Light stimuli produce dark negative afterimages, colours are complementary (red produces green and yellow produces blue) o If the primary stimulus is red, the L-cones will get tired. So if you look at the grey circle, the red-green opponent colour mechanism swings back toward the neutral point, overshoots this point and slides over to the green side. As a consequence the grey spots in the circle will appear greenish until the opponent mechanism settles back to the neutral point
  • Adaptation occurs at multiple sites in the nervous system

Does everyone see colours the same way?

  • YES:
    • People usually agree what unique green is, but there are quite some variations among different people. Some will be due to aging, which turns the lens of the eye yellow
  • NO:
    • Colour vision deficiency/colour blindness: more males than females have it because the genes that code for the L- and M-cones are on the X-chromosome
    • S-cone colour deficiencies are rare, because everyone has two copies o Deuteranope: someone who has no M-cones o Protanope: someone who has no M-cones o Tritanope: someone who has no S-cones
    • Colour-anomalous: two of the three cone photopigments are so similar that these individuals experience the world in much the same way as individuals with only two cone types: they can still make discriminations based on wavelength
    • Cone monochromat: someone who has only one cone type and thus is truly colourblind; they see the world in shades of grey
    • Rod monochromat: someone who has no cones at all and thus is truly colour-blind, has poor acuity and has serious difficulties seeing under normal daylight conditions
    • Agnosia: inability to determine what something is o Anomia: inability to name objects -MAYBE:
    • People usually agree on basic colours, but there might be some disagreement about marginal colours
    • Cultural relativism: the idea that basic perceptual experiences may be determined in part by the cultural environment

From the colour of lights to a world of colour

  • Colour contrast: a colour perception effect in which the colour of one region induces the opponent colour in the neighbouring region
  • Colour assimilation: a colour perception effect in which two colours bleed into each other, each taking on some of the chromatic quality of the other. Figure 5.21 page 138
  • Unrelated colours: colours that can be perceived in isolation
  • Related colours: colours, such as brown or grey, that can only be seen in relation to other colours
  • The colours we see in objects depend in complex ways on the colours of other objects in the vicinity

Colour constancy

  • Colour constancy: the tendency for colours to appear relatively unchanged in spite of substantial changes in lighting conditions
  • Illuminant: the light that illuminates a surface; isn’t constant
  • Spectral reflectance function: the percentage of a particular wavelength that is reflected from a surface

 

  • Spectral power distribution: the relative amount of light at different visible wavelengths

 

  • The light reflected into our eyes is the product of the surface and the illumination

 

  • However: these two products have different products of surface and illumination which are converted into two different sets of three numbers by the L- M- and S-cones. Even though the responses are very different the object is perceived as the same colour in both lighting conditions

 

Physical constraints make constancy possible

  • Colour constancy must be based on some info or assumptions that constrain the possible answers:
    • Assume that the brightest region is white, we could scale the other colours relative to this anchor. However this does not work when we are in a dark room with a red spot of light and a blue one.
    • Assumptions about the illuminant, natural light sources, and most artificial ones:
      • Are generally broadband: they contain many wavelengths
      • Their spectral composition curves, are usually smooth; spikes at a particular wavelength are uncommon
    • Assumption about surfaces:
      • Real surfaces tend to be broadband in their reflectances, even surfaces that are nearly metameric with single wavelengths of light typically reflect a wider range of wavelengths
      • The brightest thing in the visual field is likely to be white
      • A specular reflection, like the shiny spot on a billiard ball, has a wavelength composition very similar to that of the illuminant

 

  • Assumptions about the structure of the world
    • Sharp borders in an image are almost always the result of boundaries between surfaces, not boundaries between light sources
    • Shadow borders can produce a sharp edge that is unrelated to any change in the underlying surface, they are an exception to the rule. However the change across a shadow border is typically a change in brightness and not a change in the chromatic properties of the regions
  • Given the assumption that the illumination is equally distributed around the room, you will perceive strawberries as red, even in the absence of long wavelengths, because they will be reflecting more long-wavelength light than any other surface does. Figure 5.25 page 143