#!/usr/bin/env python
# -*- coding: utf-8 -*-
"""
CMCCAT2000 Chromatic Adaptation Model
=====================================
Defines CMCCAT2000 chromatic adaptation model objects:
- :class:`CMCCAT2000_InductionFactors`
- :class:`CMCCAT2000_VIEWING_CONDITIONS`
- :func:`CMCCAT2000_forward`
- :func:`CMCCAT2000_reverse`
- :func:`chromatic_adaptation_CMCCAT2000`
See Also
--------
`CMCCAT2000 Chromatic Adaptation Model IPython Notebook
<http://nbviewer.ipython.org/github/colour-science/colour-ipython/blob/master/notebooks/adaptation/cmccat2000.ipynb>`_ # noqa
References
----------
.. [1] Li, C., Luo, M. R., Rigg, B., & Hunt, R. W. G. (2002). CMC 2000
chromatic adaptation transform: CMCCAT2000. Color Research & …, 27(1),
49–58. doi:10.1002/col.10005
.. [2] Westland, S., Ripamonti, C., & Cheung, V. (2012). CMCCAT2000. In
Computational Colour Science Using MATLAB (2nd ed., pp. 83–86).
ISBN:978-0-470-66569-5
"""
from __future__ import division, unicode_literals
import numpy as np
from collections import namedtuple
from colour.adaptation import CMCCAT2000_CAT
from colour.utilities import CaseInsensitiveMapping
__author__ = 'Colour Developers'
__copyright__ = 'Copyright (C) 2013 - 2015 - Colour Developers'
__license__ = 'New BSD License - http://opensource.org/licenses/BSD-3-Clause'
__maintainer__ = 'Colour Developers'
__email__ = 'colour-science@googlegroups.com'
__status__ = 'Production'
__all__ = ['CMCCAT2000_INVERSE_CAT',
'CMCCAT2000_InductionFactors',
'CMCCAT2000_VIEWING_CONDITIONS',
'CMCCAT2000_forward',
'CMCCAT2000_reverse',
'chromatic_adaptation_CMCCAT2000']
CMCCAT2000_INVERSE_CAT = np.linalg.inv(CMCCAT2000_CAT)
"""
Inverse CMCCAT2000_CAT chromatic adaptation transform.
CMCCAT2000_INVERSE_CAT : array_like, (3, 3)
"""
[docs]class CMCCAT2000_InductionFactors(
namedtuple('CMCCAT2000_InductionFactors',
('F',))):
"""
CMCCAT2000 chromatic adaptation model induction factors.
Parameters
----------
F : numeric
:math:`F` surround condition.
"""
CMCCAT2000_VIEWING_CONDITIONS = CaseInsensitiveMapping(
{'Average': CMCCAT2000_InductionFactors(1.),
'Dim': CMCCAT2000_InductionFactors(0.8),
'Dark': CMCCAT2000_InductionFactors(0.8)})
"""
Reference CMCCAT2000 chromatic adaptation model viewing conditions.
CMCCAT2000_VIEWING_CONDITIONS : CaseInsensitiveMapping
('Average', 'Dim', 'Dark')
"""
[docs]def CMCCAT2000_forward(XYZ,
XYZ_w,
XYZ_wr,
L_A1,
L_A2,
surround=CMCCAT2000_VIEWING_CONDITIONS.get('Average')):
"""
Adapts given *CIE XYZ* colourspace stimulus from test viewing conditions
*CIE XYZ_w* colourspace matrix to reference viewing conditions *CIE XYZ_wr*
colourspace matrix using CMCCAT2000 forward chromatic adaptation model.
Parameters
----------
XYZ : array_like, (3,)
*CIE XYZ* colourspace stimulus to adapt.
XYZ_w : array_like, (3,)
Test viewing condition *CIE XYZ* colourspace matrix.
XYZ_wr : array_like, (3,)
Reference viewing condition *CIE XYZ* colourspace matrix.
L_A1 : numeric
Luminance of test adapting field :math:`L_{A1}` in :math:`cd/m^2`.
L_A2 : numeric
Luminance of reference adapting field :math:`L_{A2}` in :math:`cd/m^2`.
surround : CMCCAT2000_InductionFactors, optional
Surround viewing conditions induction factors.
Returns
-------
ndarray, (3,)
*CIE XYZ_c* colourspace matrix of the stimulus corresponding colour.
Warning
-------
The input and output domains of that definition are non standard!
Notes
-----
- Input *CIE XYZ*, *CIE XYZ_w* and *CIE XYZ_wr* colourspace matrices are
in domain [0, 100].
- Output *CIE XYZ_c* colourspace matrix is in domain [0, 100].
Examples
--------
>>> XYZ = np.array([22.48, 22.74, 8.54])
>>> XYZ_w = np.array([111.15, 100.00, 35.20])
>>> XYZ_wr = np.array([94.81, 100.00, 107.30])
>>> L_A1 = 200
>>> L_A2 = 200
>>> CMCCAT2000_forward(XYZ, XYZ_w, XYZ_wr, L_A1, L_A2) # noqa # doctest: +ELLIPSIS
array([ 19.5269832..., 23.0683396..., 24.9717522...])
"""
XYZ, XYZ_w, XYZ_wr = np.ravel(XYZ), np.ravel(XYZ_w), np.ravel(XYZ_wr)
RGB = np.dot(CMCCAT2000_CAT, XYZ.reshape(3, 1))
RGB_w = np.dot(CMCCAT2000_CAT, XYZ_w.reshape(3, 1))
RGB_wr = np.dot(CMCCAT2000_CAT, XYZ_wr.reshape(3, 1))
D = (surround.F *
(0.08 * np.log10(0.5 * (L_A1 + L_A2)) +
0.76 - 0.45 * (L_A1 - L_A2) / (L_A1 + L_A2)))
D = 0 if D < 0 else 1 if D > 1 else D
a = D * XYZ_w[1] / XYZ_wr[1]
RGB_c = RGB * (a * (RGB_wr / RGB_w) + 1 - D)
XYZ_c = np.dot(CMCCAT2000_INVERSE_CAT, RGB_c)
return np.ravel(XYZ_c)
[docs]def CMCCAT2000_reverse(XYZ_c,
XYZ_w,
XYZ_wr,
L_A1,
L_A2,
surround=CMCCAT2000_VIEWING_CONDITIONS.get('Average')):
"""
Adapts given *CIE XYZ* colourspace matrix of the stimulus corresponding
colour from reference viewing conditions *CIE XYZ_wr* colourspace matrix to
test viewing conditions *CIE XYZ_w* colourspace matrix using CMCCAT2000
reverse chromatic adaptation model.
Parameters
----------
XYZ : array_like, (3,)
*CIE XYZ* colourspace stimulus to adapt.
XYZ_w : array_like, (3,)
Test viewing condition *CIE XYZ* colourspace whitepoint matrix.
XYZ_wr : array_like, (3,)
Reference viewing condition *CIE XYZ* colourspace whitepoint matrix.
L_A1 : numeric
Luminance of test adapting field :math:`L_{A1}` in :math:`cd/m^2`.
L_A2 : numeric
Luminance of reference adapting field :math:`L_{A2}` in :math:`cd/m^2`.
surround : CMCCAT2000_InductionFactors, optional
Surround viewing conditions induction factors.
Returns
-------
ndarray, (3,)
*CIE XYZ_c* stimulus colourspace matrix.
Warning
-------
The input and output domains of that definition are non standard!
Notes
-----
- Input *CIE XYZ_c*, *CIE XYZ_w* and *CIE XYZ_wr* colourspace matrices
are in domain [0, 100].
- Output *CIE XYZ* colourspace matrix is in domain [0, 100].
Examples
--------
>>> XYZ_c = np.array([19.53, 23.07, 24.97])
>>> XYZ_w = np.array([111.15, 100.00, 35.20])
>>> XYZ_wr = np.array([94.81, 100.00, 107.30])
>>> L_A1 = 200
>>> L_A2 = 200
>>> CMCCAT2000_reverse(XYZ_c, XYZ_w, XYZ_wr, L_A1, L_A2) # noqa # doctest: +ELLIPSIS
array([ 22.4839876..., 22.7419485..., 8.5393392...])
"""
XYZ_c, XYZ_w, XYZ_wr = np.ravel(XYZ_c), np.ravel(XYZ_w), np.ravel(XYZ_wr)
RGB_c = np.dot(CMCCAT2000_CAT, XYZ_c.reshape(3, 1))
RGB_w = np.dot(CMCCAT2000_CAT, XYZ_w.reshape(3, 1))
RGB_wr = np.dot(CMCCAT2000_CAT, XYZ_wr.reshape(3, 1))
D = (surround.F *
(0.08 * np.log10(0.5 * (L_A1 + L_A2)) +
0.76 - 0.45 * (L_A1 - L_A2) / (L_A1 + L_A2)))
D = 0 if D < 0 else 1 if D > 1 else D
a = D * XYZ_w[1] / XYZ_wr[1]
RGB = RGB_c / (a * (RGB_wr / RGB_w) + 1 - D)
XYZ = np.dot(CMCCAT2000_INVERSE_CAT, RGB)
return np.ravel(XYZ)
[docs]def chromatic_adaptation_CMCCAT2000(
XYZ,
XYZ_w,
XYZ_wr,
L_A1,
L_A2,
surround=CMCCAT2000_VIEWING_CONDITIONS.get('Average'),
method='Forward'):
"""
Adapts given *CIE XYZ* colourspace matrix using given viewing conditions.
This definition is a convenient wrapper around :func:`CMCCAT2000_forward`
and :func:`CMCCAT2000_reverse`.
Parameters
----------
XYZ : array_like, (3,)
*CIE XYZ* colourspace matrix to adapt.
XYZ_w : array_like, (3,)
Source viewing condition *CIE XYZ* colourspace whitepoint matrix.
XYZ_wr : array_like, (3,)
Target viewing condition *CIE XYZ* colourspace whitepoint matrix.
L_A1 : numeric
Luminance of test adapting field :math:`L_{A1}` in :math:`cd/m^2`.
L_A2 : numeric
Luminance of reference adapting field :math:`L_{A2}` in :math:`cd/m^2`.
surround : CMCCAT2000_InductionFactors, optional
Surround viewing conditions induction factors.
method : unicode, optional
{'Forward', 'Reverse'}
Chromatic adaptation method.
Returns
-------
ndarray, (3,)
Adapted *CIE XYZ* colourspace matrix.
Warning
-------
The input and output domains of that definition are non standard!
Notes
-----
- Input *CIE XYZ*, *CIE XYZ_w* and *CIE XYZ_wr* colourspace matrices
are in domain [0, 100].
- Output *CIE XYZ* colourspace matrix is in domain [0, 100].
Examples
--------
>>> XYZ = np.array([22.48, 22.74, 8.54])
>>> XYZ_w = np.array([111.15, 100.00, 35.20])
>>> XYZ_wr = np.array([94.81, 100.00, 107.30])
>>> L_A1 = 200
>>> L_A2 = 200
>>> chromatic_adaptation_CMCCAT2000(XYZ, XYZ_w, XYZ_wr, L_A1, L_A2, method='Forward') # noqa # doctest: +ELLIPSIS
array([ 19.5269832..., 23.0683396..., 24.9717522...])
Using the CMCCAT2000 reverse model:
>>> XYZ = np.array([19.52698326, 23.0683396, 24.97175229])
>>> XYZ_w = np.array([111.15, 100.00, 35.20])
>>> XYZ_wr = np.array([94.81, 100.00, 107.30])
>>> L_A1 = 200
>>> L_A2 = 200
>>> chromatic_adaptation_CMCCAT2000(XYZ, XYZ_w, XYZ_wr, L_A1, L_A2, method='Reverse') # noqa # doctest: +ELLIPSIS
array([ 22.48, 22.74, 8.54])
"""
if method.lower() == 'forward':
return CMCCAT2000_forward(XYZ, XYZ_w, XYZ_wr, L_A1, L_A2, surround)
else:
return CMCCAT2000_reverse(XYZ, XYZ_w, XYZ_wr, L_A1, L_A2, surround)