Page 106 - Computational Colour Science Using MATLAB
P. 106

CAMS                                 93

             6.3.1  CIECAM97s

             In 1997 the CIE Technical Committee TC1-34 agreed to adopt a CAM known as
             CIECAM97s (Luo and Hunt, 1998a). CIECAM97s comprises two parts: the first
             part is a CAT that computes the corresponding tristimulus values; the second
             part calculates a set of colour-appearance descriptors for the corresponding
             tristimulus values. The particular CAT that is used in CIECAM97s is known as
             CMCCAT97 (Luo and Hunt, 1998b) and this is described in Section 6.2.2.
             CIECAM97s includes a forward and a reverse mode. The forward mode
             transforms the tristimulus values of a sample under a non-daylight illuminant,
             such as A, to those for the corresponding colour under a daylight illuminant (in
             fact, the equal-energy illuminant) and then computes some terms that describe
             the colour appearance of the sample under the daylight illuminant. The reverse
             mode is used to predict the tristimulus values under a non-daylight illuminant
             based upon the colour-appearance descriptors for a sample viewed in daylight.
               The output of the forward mode is a set of attributes that predict colour
             appearance: brightness, lightness, colourfulness, chroma, saturation and hue.
               The predictions from the CIECAM97s model are in agreement with a number
             of colour-appearance phenomena such as chromatic adaptation (CMCCAT97 is
             included within CIECAM97s), Hunt’s effect, Stevens’ effect (Stevens and
             Stevens, 1963) and the surround effect (Bartleson and Breneman, 1967).
             However, there are certain well-known colour-appearance phenomena that
             CIECAM97s cannot predict (Luo, 2002b).
               The forward mode considers a sample viewed under a (non-white) test
             illuminant against an achromatic background and computes the corresponding
             colour-appearance attributes for the sample under a (white) reference illuminant.
             The starting data required for the forward transform include the tristimulus
             values of the sample in the test illuminant (X, Y, Z), the adopted white in the test
             illuminant (X WT , Y WT , Z WT ), the reference white in the reference illuminant (X WR ,
             Y WR , Z WR ), the Y tristimulus value of the achromatic background against which
             the patch is viewed (Y ), and the luminance (in units of cd/m ) of the reference
                                                                     2
                                 B
             white (L ) and the achromatic background (L ) against which the sample is
                     W
                                                        A
             viewed.
               The model requires that values are assigned to four parameters, F, c, F LL  and
             N . Different values are recommended for these parameters depending upon the
               C
             viewing conditions (Luo, 2002b). However, for large moderately illuminated
             scenes the following values should be used: F ¼ 1.0, c ¼ 0.690, F  ¼ 0.0 and
                                                                         LL
             N ¼ 1.0. The following steps describe in detail the computation of the forward
               C
             mode of CIECAM97s.
             Step 1: Calculate the RGB values for the test sample (R, G, B) and for the
             reference white under the test (R WT , G WT , B WT ) and reference fields (R WR , G WR ,
             B WR ) using Equations (6.6). Note that this stage is equivalent to multiplying the
   101   102   103   104   105   106   107   108   109   110   111