The Eye

The human eye is a camera!I i l d l ith di l l• Iris - colored annulus with radial muscles

• Pupil - the hole (aperture) whose size is controlled by the iris• Lens - changes shape by using ciliary muscles (to focus on objects

t diff t di t )at different distances)• Retina - photoreceptor cells

Slide by Steve Seitz

Rods and cones

pigmentcone

rod

p gmolecules

Rods are responsible for intensity, cones for color perceptionRods and cones are non-uniformly distributed on the retinaRods and cones are non uniformly distributed on the retina

• Fovea - Small region (1 or 2°) at the center of the visual field containing the highest density of cones (and no rods)

Slide by Steve Seitz

Rod / Cone sensitivity

Why can’t we read in the dark?Slide by A. Efros

Electromagnetic spectrum

Human LuminanHuman Luminance Sensitivity Function

Why do we see light of these wavelengths?

© Stephen E. Palmer, 2002

…because that’s where the Sun radiates EM energy

Visible Light

Physiology of Color Vision

Three kinds of cones:N

CE

(%) 100

440

S

530 560 nm.

M L

AB

SO

RB

AN

50

RE

LATI

VE

A

400 450 500 550 600 650

WAVELENGTH (nm.)

© Stephen E. Palmer, 2002

• Ratio of L to M to S cones: approx. 10:5:1• Almost no S cones in the center of the fovea

Spectra of some real-world surfaces

metamers

The Physics of Light

Some examples of the reflectance spectra of surfaces

Wavelength (nm)

% P

hoto

ns R

efle

cted

Red

400 700

Yellow

400 700

Blue

400 700

Purple

400 700

© Stephen E. Palmer, 2002

Color spaces

• How can we represent color?

http://en.wikipedia.org/wiki/File:RGB_illumination.jpg

Color spaces: RGB

0,1,0

0,0,1

1,0,0

Image from: http://en.wikipedia.org/wiki/File:RGB_color_solid_cube.png

Some drawbacks • Strongly correlated channels • Non-perceptual

Default color space

R (G=0,B=0)

G (R=0,B=0)

B (R=0,G=0)

Linear color space CIE XYZ from RGB• Primaries are imaginary, but matching

functions are everywhere positive• The Y parameter corresponds to brightness or• The Y parameter corresponds to brightness or

luminance of a color• 2D visualization: draw (x,y), where2D visualization: draw (x,y), where

x = X/(X+Y+Z), y = Y/(X+Y+Z) Matching functions

http://en.wikipedia.org/wiki/CIE_1931_color_space

Forsyth & Ponce

Pure wavelength in chromaticity diagram

• Blue: big value of Z, therefore x and y smallx=X/(X+Y+Z) y=Y/(X+Y+Z)

Pure wavelength in chromaticity diagram

• Then y increases x=X/(X+Y+Z) y=Y/(X+Y+Z)

Pure wavelength in chromaticity diagram

• Green: y is big x=X/(X+Y+Z) y=Y/(X+Y+Z)

Pure wavelength in chromaticity diagram

• Yellow: x & y are equal x=X/(X+Y+Z) y=Y/(X+Y+Z)

Pure wavelength in chromaticity diagram

• Red: big x, but y is not null x=X/(X+Y+Z) y=Y/(X+Y+Z)

Color spaces: L*a*b* “Perceptually uniform”* color space

L (a=0,b=0)

a (L=65,b=0)

b (L=65,a=0)

(also L*u*v*)

distances quasi Euclidean

Nonlinear transformations of XYZ space.

VI. CIE COLOUR SPACES

A. CIE

CIE, the International Commission on Illumination - abbrevi-ated as CIE from its French title Commission Internationale del’Eclairage - is an organization devoted to international cooper-ation and exchange of information among its member countrieson all matters relating to the science and art of lighting [2].

In 1931 CIE laid down theCIE 1931 standard colorimetricobserver. This is the data on the ideal observer on which allcolorimetry is based [5, page 131].

B. CIE XYZ

CIE standardized theXY Z values astristimulus valuesthatcan describe any colour that can be percepted by an averagehuman observer (the CIE 1931 standard colorimetric observer).These primaries are nonreal, i.e. they cannot be realized by ac-tual colour stimuli [5, page 138]. This colour space is chosenin such a way that every perceptible visual stimulus is describedwith positiveXY Z values.

A very important attribute of theCIE XYZcolour space is thatit is device independent. Every colour space that has a trans-formation from theCIE XYZcolour space (RGB709, CIELab,CIELuv) can also be regarded as being device independent. TheCIE XYZcolour space is usually used as a reference colour spaceand is as such an intermediate device-independent colour space.

C. CIE Luv and CIE Lab colour spaces

In 1976 the CIE proposed two colour spaces (CIELuvandCIELab) whose main goal was to provide a perceptually equalspace. This means that the Euclidian distance between twocolours in theCIELuv/CIELabcolour space is strongly corre-lated with the human visual perception. To achieve this propertythere were two main constraints to take into account:• chromatic adaptation• non-linear visual response

The main difference between the two colour spaces is in thechromatic adaptation model implemented. TheCIE Labcolourspace normalizes its values by the division with the white pointwhile the CIELuv colour space normalizes its values by the sub-traction of the white point.

The transformation fromCIE XYZto CIE Luv is performedwith the following equations

L∗ = 116(

Y

Yn

) 13

− 16

u∗ = 13L∗(u′ − u′n)

v∗ = 13L∗(v′ − v′n)

for YYn

> 0.01, otherwise the following L∗ formulaeis used

L∗ = 903.3Y

Yn

Thequantitiesu′, v′ andu′n, v′n arecalculatedfrom

u′ =4X

X + 15Y + 3Z

u′n =4Xn

Xn + 15Yn + 3Zn

v′ =9Y

X + 15Y + 3Z

v′n =9Yn

Xn + 15Yn + 3Zn

ThetristimulusvaluesXn, Yn, Zn are those of the nominallywhite object-colour stimulus.

The transformation fromCIE XYZto CIE Lab is performedwith the following equations

L∗ = 116(

Y

Yn

) 13

− 16

a∗ = 500

[(X

Xn

) 13

−(

Y

Yn

) 13]

b∗ = 200

[(Y

Yn

) 13

−(

Z

Zn

) 13]

Theperceptually linear colourdifference formulaes betweentwo colours are

∆E∗ab =

√(∆L∗)2 + (∆a∗)2 + (∆b∗)2

∆E∗uv =

√(∆L∗)2 + (∆u∗)2 + (∆v∗)2

VI I . CONCLUSION

In this paper we have presented an overview of colour spacesused in image processing. We have tried to stress the importanceof the historical and perceptual background that has led to theintroduction of these colour spaces.

REFERENCES

[1] Symon D’O. Cotton,Colour, colour spaces and the human visual system,School of Computer Science, University of Birmingham, England, Techni-cal Report, B15-2TT, May 1996.

[2] CIE home pagehttp://members.eunet.at/cie/ .[3] Charles Poynton,A Guided Tour of Color Space, New Foundations for

Video Technology (Proceedings of the SMTPE Advanced Television andElectronic Imaging Conference), 1995, pages 167-180.

[4] Charles Poynton, Frequently Asked Questions about Color,http://www.inforamp.net/ poynton, 1999.

[5] Gunter Wyszecki, W.S. Stiles,Color Science Concepts and Methods, Quan-titative Data and Formulae, John Wiley and Sons, Inc, 2000.

[6] Mark D. Fairchild,Color Appearance Models, Addison Wesley, 1998.[7] Henryk Palus,Colour spaces, Chapmann and Hall, 1998.[8] Adrian Ford and Alan Roberts,Colour space conversions, Westminster

University, London, 1998.

L between 0 and 100u* between -134 and 200v* between -140 and 122

L*u*v* theinner solid

Lab

Only color shown – constant intensity

If you chose only chrominance (say, a and b)...

Only intensity shown – constant color

... if you chose only luminance (say, L).

Most information in intensity.

Original image

http://colorcalc.gc-taylor.com/

Several distances between two (Lab) color exist. The CIE2000is used today in many places.Adds five corrections for the huerotation, neutral colors, lightness,chrome and hue. The difference is complicated but better thanCIE76, which is similar to the one used in mean shift.Function of that color is checked, the perceived minimumdifference still can change!