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

Transcription

1 Color

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

3 Color Science Light & Spectra Light is an electromagnetic wave It s color is characterized by its wavelength Laser consists of single wavelength Most light sources produce contributions over many wavelengths, contributions that fall in the visible wavelength can be seen λ Vs Spectral power curve is called spectral power distribution E(λ) (SPD). Light from 400 to 700 nanometer (10-9 meter) Computer Graphics 3

4 Color Science Light & Spectra [2] Computer Graphics 4

5 Color Science Light & Spectra [3] Red light has longer wavelength in the visible light Blue light has the shorter. The shorter the wavelength, higher h the vibration & energy Red photons carry around 1.8eV & blue 3.1eV Computer Graphics 5

6 Color Science Vision & Sensitivity Eye resembles a camera focusing image into the retina Retina consists of rods & 3 kind of cones Rods have more sensitivity when the light level lis low & produce image in gray The higher the light level, the more the sensitivity with the cones & more neurons firing The 3 kinds of cones are more sensitive to R, G & B present in the ratios 40:20:1 Computer Graphics 6

7 Color Science Vision & Sensitivity [2] Eye is most sensitive to the middle of the visible spectrum Let us denote the spectral sensitivity of R, G, B cones as a vector q(λ) q(λ)=[q R (λ), q G (λ), q B (λ)] T Computer Graphics 7

8 Color Science Vision & Sensitivity [3] The sensitivity of each cones can be specified as R= E(λ) q R (λ) d λ (1) G= E(λ) q G (λ) d λ (2) B= E(λ) q B (λ) d λ (3) - integral Equations 1, 2, 3 quantify the signals transmitted to the brain Computer Graphics 8

9 Color Science Image Formation Equations 1,2 & 3 applies for self luminous objects, but we mostly receive image light reflected from objects The reflectance S(λ) varies from object to object The formations would be as follows Light from illuminant with SPD E(λ) falls on the object with reflectance S(λ) and is reflected, then filtered by the eyes sensitivity functions q(λ) & the image formation model is now R= E(λ) S(λ) q R (λ) d λ (4) G= E(λ) S(λ) q G (λ) dλ λ (5) B= E(λ) S(λ) q B (λ) d λ (6) - integral Computer Graphics 9

10 Color Science Image Formation[2] The main characteristics of a color for human perception are the following ones. hue corresponding to the dominant wavelength in the spectrum of the color saturation or purity which is high when the spectrum consists of a narrow peak at the dominant wavelength, and which is low for a flatter spectrum. intensity or lightness depending on the energy of the spectrum Computer Graphics 10

11 Color Science Gamma Correction The RGB in the image files are converted to analog & drive the electron guns of CRT The light emitted from CRT is not linear to the voltage applied, it is proportional to voltage raised to the power, called gamma This effect can be precorrected by applying inverse transformation SMPTE( Society of Motion Picture and Television Engineers) set this value to 2.2 Computer Graphics 11

12 Color Science Gamma Correction [2] From digital Image Processing C.Gonzalez & R.Woods Computer Graphics 12

13 Color Science CIE Chromaticity Diagram In 1931,CIE (Commission i Internationale de L Eclairage) combined all results of color matching experiments in the form of CIE color matching functions The reason why some parts of the color matching functions curve are negative? Shift color primaries to the other side to produce those colors CIE color matching functions Computer Graphics 13

14 Color Science CIE Chromaticity Diagram [2] The CIE decided to use imaginary lights that had two especially nice features: one of the lights is gray and provides no hue information; the other two lights have zero luminance and provide only hue The response for these three lights is defined d by the triple (X, Y, Z) where Y is the luminance ( / brightness/ perceived intensity) Computer Graphics 14

15 (λ), (λ), (λ) Color Science CIE Chromaticity Diagram [3] The formula for the CIE tristimulus values (X, Y, Z) is X= E(λ) x(λ) d λ (7) Y= E(λ) y(λ) d λ (8) Z= E(λ) z(λ) d λ (9) The values of are obtained from the original r, b, g curves, by multiplication by a 3x3 matrix. CIE standard color matching functions Computer Graphics 15

16 Color Science CIE Chromaticity Diagram [4] Factoring out luminance to concentrate on color, we get x= X /(X+Y+Z) y= Y /(X+Y+Z) z= Z /(X+Y+Z) Now z=1-x-y, means z is redundant Plotting x vs. y for colors in the visible spectrum we get CIE Chromaticity diagram Computer Graphics 16

17 Color Science CIE Chromaticity Diagram [5] The Line Joining i red to violet is called purple line, and is not a part of the spectrum Interior points specify all the visible color combinations Standard approximations of day light by CIE includes illuminant C ( , ) D65 ( , ) D100 (0.2788, ) F2 (for fluorescent illumination) Computer Graphics 17

18 Color Science CIE Chromaticity Diagram [6] So can you get X, Y & Z from x, y & z? No. For complete description of color, we need x, y & Y. So, x, y represent chromaticity & Y represent luminance Computer Graphics 18

19 Color Science Color Monitor Specifications When R, G, B electron guns are fired at the maximum voltage (normalized to [0 1]), white point should appear. Otherwise use the voltage that generates the white point. Color monitor specifications includes the white point chromaticity desired Chromaticities & white point of monitor specifications Computer Graphics 19

20 Color Science Out-of-Gamut colors Gamut refers to the subset of colors which can be accurately represented in a given color space or by a certain output device Colors are said to be out of gamut, if it is perceivable by the user & is not representable on the device being used Given x, y (& z is implied) the R, G, B can be found from Computer Graphics 20

21 Color Science Out-of-Gamut colors [2] The color is out-ofgamut, if from the previous equation, any of R, G or B turns negative Approximate with the closest in-gamut color Computer Graphics 21

22 Color Science Munsell Color Naming System Color space is continuous, naming of colors requires quantization A common color space used for naming is HSI color space. Each of the values can gradually change. Gradual change of 'hue' can be described by names such as red, orange red, orange, yellowish orange, yellow, green yellow, green, sea green, cyan, blue, violet and purple. Computer Graphics 22

23 Color Science Munsell Color Naming System [2] Gradual change of saturation may for example be expressed by adding to a color name labels like gray, grayish, moderate, strong and vivid. Gradual change of intensity may be expressed by adding labels such as 'blackish', very dark, dark, medium, light (or pale) and very light. Munsell Naming System uses three axes to naming colors Value (black to white) (9 steps) Hue (40 steps around a circle) *Radius of circle varies Chroma (16 levels) This is called the CNS model. Computer Graphics 23

24 Color Science Other Color Coordinate Systems CMY (Cyan-Magenta-Yellow) YIQ (Luminance, In-phase, Quadrature) -It is used by the NTSC television norm. HSL(Hue-Saturation-Lightness) HSV(Hue-Saturation-Value) Saturation Value) HSI(Hue-Saturation-Intensity) HCI(Hue-Chroma-Intensity) HVC(Hue-Value-Chroma) HSD(Hue-Saturation-Darkness) Computer Graphics 24

25 The YIQ model It is not based in the 3 primary colors, but Y represents the luminance I and Q represent the in-phase and quadrature values when quadrature modulation is used to specify the crominance C = Icos(F sc t) + Qsin(F sc t) where F sc is the frequency 3.58MHz Computer Graphics 25

26 The YIQ model [2] It can be obtained from the R, B, G model using the transformation It allows easy adjustment of the luminance for all monitors Computer Graphics 26

27 The HSV model As the name suggests, this model is based on the hue, saturation and value(of intensity). Computer Graphics 27

28 The HSV model[2] d l The values are obtained by: max is the greatest of r, g, and b, and min the least (coordinates values for R, G and B). Computer Graphics 28

29 The HSL model As the name suggests, this model is based on the hue, saturation and lightness. The hue is obtained as in the HSV model The s and l (between 0 and 1) by Computer Graphics 29

30 Color Models for Image RGB Vs CMY Additive Vs Subtractive Models Additive model Used in computer displays, Uses light to display color, Colors result from transmitted light Red+Green+Blue=White Subtractive Models Used in printed materials, Uses ink to display color, Colors result from reflected light Cyan+Magenta+Yellow=Black Computer Graphics 30

31 Color Models for Image RGB Vs CMY [2] Computer Graphics 31

32 Color Models for Image RGB Vs CMY [3] RGB & CMY Cubes Computer Graphics 32

33 Color Models for Image RGB Vs CMY [4] Conversion From RGB to CMY Conversion From CMY to RGB Computer Graphics 33

34 Color Models for Image CMYK Eliminating i i amounts of yellow, magenta, and cyan that would have added to a dark neutral (black) and replacing them with black ink Four-color printing uses black ink(k) in addition to the subtractive primaries yellow, magenta, and cyan. Reasons for Black addition includes CMY Mixture rarely produces pure black Text is typically printed in black and includes fine detail Cost saving : Unit amount of black ink rather than three unit amounts of CMY Computer Graphics 34

35 Color Models for Image CMYK [2] The transformation from CMY to CMYK is given by the formulas: K = min(c, M, Y) C = C K M = M K Y = Y K Note that this implies that at least one of the four values will be zero. Computer Graphics 35

36 Color Models for Image CMYK[2] Used especially in the printing of images = Computer Graphics 36