Colour Vision

1
Colour Vision the Rail Industry

• Dr Catharine Chisholm
• c.m.chisholm_at_bradford.ac.uk

2
Acknowledgements

• Rail Standards Safety Board
• Dr Richard Graveling, Institute of Occupational
  Medicine

3
Background

• Optometrist
• Lecturer in Optometry and Vision Science
• Anglia, City, Bradford
• PhD
• Effect of scattered light on visual performance
  post laser refractive surgery
• Occupational vision research
• Evidence-based visual standards
• CAA, Home Office, DfT, HMPS, NATS, RSSB

4
Presentation outline

• Current colour vision standards
• Examination of safety critical colour tasks
• Background on colour vision
• The assessment of colour vision
• Acquired colour defects
• Relevance of colour tests to the rail industry

5
Current colour vision requirements

• EU directive (2007/59/EC) Annex II
• Normal colour vision use of a recognised test,
  such as Ishihara, as well as another recognised
  test if required 
• Recognition of colour signals the test shall be
  based on recognition of single colours and not on
  relative differences

6
What does this mean in practice?

• The vast majority of colour normals pass
• Only the mildest red/green colour defectives fail
• The tiny proportion of the population with blue
  /yellow defects will pass
• Not examined by Ishihara
• 1 in 13,000
• Current practice is safe
• DDA?

7
Visual task analysis

• RSSB Human Factors (RSSB, 2011)
• Interviews
• Site visits
• Discussions with equipment suppliers
• Areas of focus
• Trackside signals
• Indicator lights and displays
• Electrical and telecommunications wiring

8
Safety critical colour tasks

• Connotative - no redundancy related to positional
  cues or labelling
• Serious consequences of failure
• Spectral characteristics must be well defined
• Hardware must meet standards across the network
  and over time

9
Safety critical task for train drivers and others

• Colour light signals
• Red
• Yellow
• Green
• Defined in Railway Group Standard GK/RT0045
• Specified in BS 1376 1974

10
Colour light signals

• Train drivers
• Alerted to trackside signals by track-based alarm
  and AWS sunflower
• Nevertheless, rapid identification may be needed
  so vision remains important
• In addition, some trackside workers and others
  need to identify head and tail lights on trains
  and vehicles, to identify direction of travel
• Red
• White Specified in BS 1376, 1974

11
Environmental conditions

• Conditions of reduced visibility
• Remove positional cues
• Reduce brightness (sighting distances)
• Increase scattering of short wavelengths
• Green may look yellow
• Green, yellow, white will NOT be confused with
  red by a colour normal

12

• Any colour can be specified in terms of x,y
  coordinates on the CIE chromaticity diagram

13
Reading repeater indicators / diagnostic LEDs /
other indicators

• Train drivers, signallers, level crossing
  attendants, telecommunications workers etc.

• Train cabs
• All colour redundant
• Positional and labelling cues
• Back-up alarms

14
Signal control centres

• Lever controls

15
Signal control centres

• Lever controls signal status lights
• No colour dependencies for
  lever-controlled signal

16
Signal control centres

• NX mimic displays
• Colour redundant other than signal status light
  (red / green)

17
NX signal status lights

• No definition of colours to be used
• No consistency between sites
• At discretion of manufacturer
• Lamps and lenses discolour with age

18
Signal control centres

• Computer-based mimic displays - complex scene.
  Poor use of colour
• Route sectors magenta /grey
• Lack of standardisation of specific colours

19
Signallers - summary

• Lever control colour not safety critical
• Also, Bardic lamp and flags can be labelled
• NX mimic board signal status lights
• No standardisation of colours used
• Computer-based mimic route sectors
• No standardisation of colours used
• Different categories of signaller?
• Logistics?
• Movement of individuals?

20
Distinguishing between different colour cables

• Signalling and telecoms engineers
• Wiring is colour coded
• red, blue, yellow, green, slate, orange, white,
  purple, and black (telecommunications and signal
  wires)
• Other codes also in use (RAL codes)
• standard electrical colours (green/yellow, blue,
  brown)
• Older electrical colours (red and black)

21
Distinguishing between different colour cables

• No colour redundancy
• Range of colour codes used by different
  manufacturers of equipment
• Colours not specified and may vary from batch to
  batch
• Not possible to define colour requirements for
  these roles

22
Colour Vision
Colour is the visual effect caused by the
spectral composition of light emitted,
transmitted, or reflected by objects

• The colour appearance of an object depends on
• The colour properties of the light source
• The absorption/reflection properties of the
  object
• The colour processing of the eye and brain

23
The cones
lt7 blue cones at fovea

• 6 million
• 3 types (40201 RGB)
• Concentrated at the fovea
• Less sensitive than the rods
• Good visual acuity
• Excellent colour vision
• Discriminate 7 million colours and shades

Better colour discrimination foveally
24
Normal Colour Vision

• Normal Trichromacy All spectral hues can
  be matched with a mixture of 3 primaries, long
  wavelength (red), middle wavelength (green) and
  short wavelength (blue) lights. (L, M, and S
  sensitive cones)
• Red - Green colour vision is stable with age
• Short wavelength (blue) sensitivity varies
  reduces with age (gt55 years)

25
(No Transcript)
26
Signal Yellow
27
Signal green
28
Colour signal transmission
29
Higher colour processing
30
C.I.E Chromaticity Diagram 1931MacAdam ellipses
All colours within an individual ellipse look
identical to a normal trichromat
31
Colour defectives

• Congenital (gene mutation) or acquired (damage to
  retina or visual pathway)
• Cannot appreciate or discriminate between as many
  colours as a normal person
• Colours which look different to people with
  normal colour vision, look the same and are
  confused (if there is no perceived lightness
  difference). Poor discrimination of specific
  hues.

32
Congenital colour deficiency

• Present from birth and of constant severity and
  type
• Same defect in each eye
• More common in males than females
• Red/green defects far more common than blue
• Range in severity from very mild to colour
  blind (monochromats)
• Clinical tests are designed to look for common
  colour confusions associated with congenital
  defects
• Normal visual function (except monochromats)

33
Congenital colour deficiency

• 8 men
• 0.5 women

• Confuse colours of the same brightness
• Distinguish fewer colours
• Either one or more cone types
• missing (dichromat) or

• deficient (anomalous trichromat)

34
Anomalous trichromat 1 deficient cone type
Dichromat 1 cone type absent
Protanope 1
Severe
Variable mild to severe
Protanomalous 1
Deuteranope 1
Severe
Variable mild to severe
Deuteranomalous 5
35

• Congenital tritan defects (blue/yellow)

• Just as common in males as females
• The exact prevalence is unknown
• Tritanopia 1 in 13,000
• Tritanomalous trichomatism 1 in 500?

36
Complete Colour Blindness Monochromatism /
Congenital Achromatopsia

• Rod Monochromats
• No functioning cones
• Vision very poor
• Equal malefemale
• Prevalence 1 in 35,000

• Cone Monochromats
• Single cone response
• short-wave (blue) cones only
• Hue discrimination in low light
• Vision below average to poor
• Extremely rare

37
Which colours are confused?
38
Which colours are confused?
39
Colour confusions for anomalous trichromats

• Same colours confused as for dichromats but only
  more similar colours, closer together on the CIE
  diagram.
• Confusion ellipses vary in size and include a
  range of desaturated colours depending on the
  severity of the colour defect
• Severe anomalous trichromatism
• isochromatic zones (as for dichromats)
• Mild anomalous trichromatism
• ellipses only slightly extended compared to
  normal trichromats

40
Isochromatic confusion lines
41
Tritan Isochromatic zones
42
Colour confusions
43
How colour deficients see
William Stroudleys painting of a steam engine
improved engine green
44
Sensitivity to red lights

• Those without a L (red) cone have no retinal
  receptor that can respond to the far red end of
  the spectrum
• Protanopes dont see the far red
• Milder protanomalous - red seen as dim
• I find the intensity of the red light
  diminished. If I am a distance away from the
  light, the amplitude of light is too low for me
  to see at all and thus I can't determine if it is
  the top, middle or bottom light burning. There
  just are no lights! As I get closer to the light
  the intensity increases and I start to see the
  red light. To compensate I watch the other
  vehicles ahead and if they start to reduce speed
  I know there must be a red light ahead.

45
Colour Vision Testing

• Screening - Abnormal / normal
• Ishihara (red-green), Hardy-Rand-Rittler
• Grading - Mild defects pass, severe defects fail
• Farnsworth D15, City Uni 2nd Ed.
• Classification protan, deutan or tritan
• Most tests although some better than others
• Vocational pass/fail criteria based on task
  analysis
• Practical colour matching
• Colour recognition (lanterns)

46
Illumination

• CIE Standard Source C (6700K) or Source D65
  (6500K)
• MacBeth Easel lamp (350-400lux)
• Corrected daylight fluorescent
• Appearance of pigment colours changes with the
  spectral content of the illuminant
• Tungsten and fluorescent alter the appearance of
  colours
• Some slight deuteranomalous trichromats will pass
  the test

47
Ishihara Pseudoisochromatic plates

• Screening/identifying red-green colour deficiency
  only
• Exploits isochromatic colour confusions with
  colour camouflage and luminance masking
• Full version 38 plates, 25 numerals and 13
  pathways, 24 plate common
• Viewing at 66cm distance, 4 secs for each plate.
  Lighting critical
• Randomise plates
• Do not touch the plates
• Protect from light and dust

48
Ishihara 24 plate pass/fail

• Fail 1 or more on plates 2-15
• 15 of normals fail
• Fail 2 or more on plates 2-15
• Sensitivity
• Fails 99 of colour defectives
• Specificity
• Passes 94 of colour normals (Vingrys 1984) but
  100 if examiner allows for misreadings (Birch
  2001)

49
Ishihara 24 plate

• Plate 1 demonstration
• Plates 2-15 Transformation or vanishing digits
• Do not use hidden digit (poor sensitivity)
• Plates 16-17 classification plates
• Correctly classifies as protan or deutan
  only 50 of the time

50
Hardy-Rand-Rittler

• 4th edition Richmond HRR (2002)
• 4 demonstration plates
• 6 screening plates
• 4 red/green
• 2 blue/yellow
• 14 classification plates
• 10 red/green
• 4 blue/yellow
• classify severity and type

51
Performance of 4th Ed. HRR

• Fail two or more errors on the six screening
  plates
• Sensitivity 100
• Specificity 97.5
• Fail three or more errors from the six screening
  plates
• Sensitivity 98
• Specificity rises to 100 (no misreadings)
• Performance for detecting tritan defects is
  unknown

52
Vision Screeners
53
Testing procedure

• Ishihara and HRR both provide a good initial
  screening
• Secondary tests
• Should always be preceded by screening
• Should relate to visual task
• Are used in optometric practice for
  classification or type and/or severity
• City University, Farnsworth D15, Lantern tests,
  CAD test

54
The City University test (2nd edition)

• 10 plates central colour and 4 peripheral
  colours
• Isochromatic colour confusions for protan,
  deutan, tritan
• Which of the 4 outer colours looks most like the
  colour in the centre?
• Lighting critical
• Only 3rd Ed available unvalidated
• Aims to screen and grade severity but poor design

55
The City University test (2nd edition)

• Fail 1 or more errors (Birch 1997)
• Fail 2 or 3 errors allows some protans to pass
• Passes mild colour defectives
• Fails 97 of dichromats and 36 of anomalous
  trichromats
• gt60 of people who fail obtain mixed
  protan/deutan classification results

56
Farnsworth D15

• Designed for vocational guidance (e.g.
  electrical)
• Hue discrimination task

• Put the colours back into the box in a natural
  hue order
• Lighting critical
• Easy to appreciate what colours the subject
  confuses

57
Farnsworth D15

• People with mild colour deficiency pass this test
• Similar sensitivity to City Uni 2nd Ed.
• Fail 2 or more crossings of the circle
• Repeat all fails
• Grading and classifying protan, deutan and tritan
  deficiency

58
The Farnsworth-Munsell 100 hue test

• Examines hue discrimination ability
• Not reliable for screening but OK for grading
  (moderate to severe colour deficiency)

59
The Nagel Anomaloscope

• Gold standard (for RG)
• identification and diagnosis of RG deficiencies
• Red (670nm) Green (546nm) Yellow (589nm)

60
(No Transcript)
61
Nagel Anomaloscope

• Can distinguish dichromats from anomalous
  trichromats
• Allows you to identify protans with impaired
  sensitivity to red
• Accurately distinguishes protans from deutans
• Expensive and only available at specialist centres

62
Lantern tests

• Distance viewing
• Involve naming of coloured lights
• Some relate to actual tasks
• Australian RailCorp lantern

• Others designed to detect colour deficiency
• CAM lantern, CNLAN, Farnsworth, Optec900
• Name Red Green White Yellow, Blue
• Number of errors does not grade severity
• Different lanterns produce highly variable
  outcomes (Squire et al. 2005)

63
The CAD test

• Computer-based test
• Quick, initial screening
• Longer test plots colour discrimination ellipse
  in all colour directions
• Can be compared to customised limits relating to
  specific colour task
• CAD used by Civil Aviation Authority and London
  Underground

64
(No Transcript)
65
X Chrom / ChromaGen Lenses

• EU directive (2007/59/EC)
• Coloured contact lenses and photochromatic lenses
  are not allowed. Sunglasses with specific BSEN
  tint are allowed

• Coloured filters change colour appearance by
  subtracting some wavelengths and altering
  relative colour lightnesses
• May pass isochromatic tests but colour defect
  shifted to another axis
• Monocularly worn filters compromise binocular
  vision

66
Acquired colour deficiency

• Cannot appreciate or discriminate between as many
  colours as a normal person
• Onset after birth
• Changes in type and severity over time
• E.g. R/G then B/Y then colour blind
• Deficiency likely to be different in each eye
• As common in males as females
• May affect occupational suitability

67
Common causes

• Eye diseases
• Cataract
• Glaucoma
• ARMD
• Optic neuritis
• Systemic
• Diabetes
• Blue-Yellow defect
• May be present before retinal changes

68
Drug related

• Quinine, chloroquine
• Ethambutol
• Digoxin
• Tobacco
• Alcohol

69
Which tests are relevant to the rail industry?

• Safety critical tasks
• Trackside signals
• Vehicle and train lamps
• Range of colour vision tests selected based on
  known performance, ease of use, practicalities
• Ishihara, HRR
• City 2nd Edition, D15
• CNLAN, RailCorp lantern, CAM lantern

70
Which tests are relevant to the rail industry?

• Comparison of performance between mock-up and
  each test, or test combination
• 40 colour defectives
• 40 colour normals
• There may not be a suitable combination
• Test customised to particular colour tasks?

71
Thank you!