Colour Vision

• The Perceptual Dimensions of Colour: HSB

  • colour is a normal distribution (bell curve)

  1)HUE = the mean of the distribution of wavelengths entering the eye

  *what we define as colour

  2)SATURATION = the variance / width of the distribution of wavelengths

  *how washed out a colour is

  *the higher the peak (no. of photons) the more saturated that wavelength of light is

  3)BRIGHTNESS = the area under the curve

  *how light a colour is - linked w/ saturation as the higher the area, the higher the saturation

  
• How Many Colour Can Humans Distinguish? 200 20 500

  • Visible spectrum = 400 - 720 nm
  • Hue = 200 steps
  • Saturation = 20 steps
  • Brightness = 500 steps
  • Total Steps = 2,000,000

  ! Artists can distinguish more colours

  • Colour vision arises from cone photoreceptors

  *S cones = short wavelength sensitive (BLUE)

  *M cones = medium wavelength sensitive (GREEN)

  *L cones = long wavelength sensitive (RED)
• Emission/Illumination Spectrum of Light Sources:

  • Most light sources emit light at MANY wavelengths, but some more wavelengths more than others

  *Tungsten Emission of Red > Sunlight Emission of Red

  *Tungsten Light = more red light emission

  *Fluorescent Light = more green light emission

  Reflectance Spectrum of Objects:

  • Different objects reflect different amount of the various wavelengths of light falling on them  

  *We will see highest peak at whatever colour the object is
• Colour Constancy:

  • The ability to perceive the reflectance properties of surfaces despite changes in illumination
  • The luminance spectrum we see is the product of the reflectance spectrum and the illuminance  spectrum [ L = R X I ]
  • DAYLIGHT: I spectrum is uniform, L spectrum is similar to R spectrum
  • TUNGSTEN: more difficult to work out R spectrum – things will look more red due to I spectrum tending to red wavelengths
  • MONOCHROMATIC SOURCE: we will only view object in light of I spectrum so impossible determine R spectrum
• The Retinex Theory of Colour Constancy:

  • (Retina and Cortex) visual system compares the luminance profile from adjacent surfaces/regions and determines whether the changes in colour are due to changes in illumination or genuine changes in colour

  *Changes in colour due to illumination = large areas of a scene

  *Changes in colour due to the actual colour = local area of a scene

  • Regarded as the most likely mechanism of colour constancy
• How can Colour Constancy be fooled?

  • Light and shadows - can fool us into thinking that 2 objects are a different colour
  • Monochromatic light source illumination - can make meat look more fresh as it's only viewed under a single wavelength
• 1)Young-Helmholtz Trichromatic Theory of Colour Vision:

  • the human eye contains three primary colour receptors that produce red, green, and  blue
  • Each colour is represented by a unique combination of the outputs of the 3 colour receptors.

  GOOD:

  • Color matching - where a combination of the 3 primaries can match a colour
  • Some findings from studies of colour blind patients.

  BAD:

  • Colour experiences are always lost in pairs: red/green , blue/yellow
  • Colour afterimages – when staring at one colour the​

  afterimage is the opposite colour.​

  • Colour mixtures – only certain ones can mix
• 2)Opponent Colour Theory of Colour Vision:

  • 4 primaries organised into 2 opponent chromatic channels - red&green, blue&yellow
  • A third achromatic channel - black&white

  GOOD:

  • The primacy of red, green, blue and yellow.​
  • Opposite coloured afterimages - blue has yellow afterimage
  • Duality of loss in colour blind patients instead of a single colour

  BAD:

  • Each photoreceptor is sensitive to red , green or blue - NOT SOMEWHERE IN BETWEEN
• 3)Dual Process Colour Theory of Colour Vision:

  • Proved that both theories were correct but for different stages of  colour processing.​

  *First Stage of Processing = Trichromatic Theory

  *Second Stage of Processing = Opponent Theory

  ! All theories were based purely on behavioural data/theory as adaptive optics did not exist yet

  ! Adaptive optics enabled us to image retina and find there were 3 types of cone photoreceptors - arrangement and no. of each cone type was unique to the individual
• How is Colour-Blindness Diagnosed?

  • Ishihara Colour Plates - px asked to detect no. on the plate
  • Farnsworth 100 Hue Test - for each of the 4 hues ( red , green , blue , yellow ) px orders 25 pieces within each hue as they see fit
  • Adaptive Optics - there are missing cones corresponding to where the third cone type would have been (density reduced)

  *this does not affect VA , unless a substantial amount of cones are missing

  *before adaptive optics, it was thought that the cones that px DID have would increase to compensate for the missing cones
• Colour Blindness/Colour-Vision Defects: PDT

  • affects 8% of males , <1% females
  • Normal Colour Vision = trichromat (use all 3 cone types)
  • Colour Blind = dichromat (only have 2 cone types)

  DICHROMACY:

  1)PROTANANOPIA = missing red cones , causing red-green confusion

  2)DEUTERANOPIA = missing green cones , causing green-red confusion

  3)TRITANOPIA (rare) = missing blue cones , blue-yellow confusion

  Anomalous Trichromacy:

  • not as severe as dichromacy

  *PROTANANOMALY = altered red cones

  *DEUTERANOMALY = altered green cones

  *TRITANOMALY = altered blue cones