1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
|
# -*- coding: utf-8 -*-
# Copyright (c) Vispy Development Team. All Rights Reserved.
# Distributed under the (new) BSD License. See LICENSE.txt for more info.
from __future__ import division # just to be safe...
import numpy as np
###############################################################################
# Utility functions
def _check_color_dim(val):
"""Ensure val is Nx(n_col), usually Nx3"""
val = np.atleast_2d(val)
if val.shape[1] not in (3, 4):
raise RuntimeError('Value must have second dimension of size 3 or 4')
return val, val.shape[1]
###############################################################################
# RGB<->HEX conversion
def _hex_to_rgba(hexs):
"""Convert hex to rgba, permitting alpha values in hex"""
hexs = np.atleast_1d(np.array(hexs, '|U9'))
out = np.ones((len(hexs), 4), np.float32)
for hi, h in enumerate(hexs):
assert isinstance(h, str)
off = 1 if h[0] == '#' else 0
assert len(h) in (6+off, 8+off)
e = (len(h)-off) // 2
out[hi, :e] = [int(h[i:i+2], 16) / 255.
for i in range(off, len(h), 2)]
return out
def _rgb_to_hex(rgbs):
"""Convert rgb to hex triplet"""
rgbs, n_dim = _check_color_dim(rgbs)
return np.array(['#%02x%02x%02x' % tuple((255*rgb[:3]).astype(np.uint8))
for rgb in rgbs], '|U7')
###############################################################################
# RGB<->HSV conversion
def _rgb_to_hsv(rgbs):
"""Convert Nx3 or Nx4 rgb to hsv"""
rgbs, n_dim = _check_color_dim(rgbs)
hsvs = list()
for rgb in rgbs:
rgb = rgb[:3] # don't use alpha here
idx = np.argmax(rgb)
val = rgb[idx]
c = val - np.min(rgb)
if c == 0:
hue = 0
sat = 0
else:
if idx == 0: # R == max
hue = ((rgb[1] - rgb[2]) / c) % 6
elif idx == 1: # G == max
hue = (rgb[2] - rgb[0]) / c + 2
else: # B == max
hue = (rgb[0] - rgb[1]) / c + 4
hue *= 60
sat = c / val
hsv = [hue, sat, val]
hsvs.append(hsv)
hsvs = np.array(hsvs, dtype=np.float32)
if n_dim == 4:
hsvs = np.concatenate((hsvs, rgbs[:, 3]), axis=1)
return hsvs
def _hsv_to_rgb(hsvs):
"""Convert Nx3 or Nx4 hsv to rgb"""
hsvs, n_dim = _check_color_dim(hsvs)
# In principle, we *might* be able to vectorize this, but might as well
# wait until a compelling use case appears
rgbs = list()
for hsv in hsvs:
c = hsv[1] * hsv[2]
m = hsv[2] - c
hp = hsv[0] / 60
x = c * (1 - abs(hp % 2 - 1))
if 0 <= hp < 1:
r, g, b = c, x, 0
elif hp < 2:
r, g, b = x, c, 0
elif hp < 3:
r, g, b = 0, c, x
elif hp < 4:
r, g, b = 0, x, c
elif hp < 5:
r, g, b = x, 0, c
else:
r, g, b = c, 0, x
rgb = [r + m, g + m, b + m]
rgbs.append(rgb)
rgbs = np.array(rgbs, dtype=np.float32)
if n_dim == 4:
rgbs = np.concatenate((rgbs, hsvs[:, 3]), axis=1)
return rgbs
###############################################################################
# RGB<->CIELab conversion
# These numbers are adapted from MIT-licensed MATLAB code for
# Lab<->RGB conversion. They provide an XYZ<->RGB conversion matrices,
# w/D65 white point normalization built in.
# _rgb2xyz = np.array([[0.412453, 0.357580, 0.180423],
# [0.212671, 0.715160, 0.072169],
# [0.019334, 0.119193, 0.950227]])
# _white_norm = np.array([0.950456, 1.0, 1.088754])
# _rgb2xyz /= _white_norm[:, np.newaxis]
# _rgb2xyz_norm = _rgb2xyz.T
_rgb2xyz_norm = np.array([[0.43395276, 0.212671, 0.01775791],
[0.37621941, 0.71516, 0.10947652],
[0.18982783, 0.072169, 0.87276557]])
# _xyz2rgb = np.array([[3.240479, -1.537150, -0.498535],
# [-0.969256, 1.875992, 0.041556],
# [0.055648, -0.204043, 1.057311]])
# _white_norm = np.array([0.950456, 1., 1.088754])
# _xyz2rgb *= _white_norm[np.newaxis, :]
_xyz2rgb_norm = np.array([[3.07993271, -1.53715, -0.54278198],
[-0.92123518, 1.875992, 0.04524426],
[0.05289098, -0.204043, 1.15115158]])
def _rgb_to_lab(rgbs):
rgbs, n_dim = _check_color_dim(rgbs)
# convert RGB->XYZ
xyz = rgbs[:, :3].copy() # a misnomer for now but will end up being XYZ
over = xyz > 0.04045
xyz[over] = ((xyz[over] + 0.055) / 1.055) ** 2.4
xyz[~over] /= 12.92
xyz = np.dot(xyz, _rgb2xyz_norm)
over = xyz > 0.008856
xyz[over] = xyz[over] ** (1. / 3.)
xyz[~over] = 7.787 * xyz[~over] + 0.13793103448275862
# Convert XYZ->LAB
L = (116. * xyz[:, 1]) - 16
a = 500 * (xyz[:, 0] - xyz[:, 1])
b = 200 * (xyz[:, 1] - xyz[:, 2])
labs = [L, a, b]
# Append alpha if necessary
if n_dim == 4:
labs.append(np.atleast1d(rgbs[:, 3]))
labs = np.array(labs, order='F').T # Becomes 'C' order b/c of .T
return labs
def _lab_to_rgb(labs):
"""Convert Nx3 or Nx4 lab to rgb"""
# adapted from BSD-licensed work in MATLAB by Mark Ruzon
# Based on ITU-R Recommendation BT.709 using the D65
labs, n_dim = _check_color_dim(labs)
# Convert Lab->XYZ (silly indexing used to preserve dimensionality)
y = (labs[:, 0] + 16.) / 116.
x = (labs[:, 1] / 500.) + y
z = y - (labs[:, 2] / 200.)
xyz = np.concatenate(([x], [y], [z])) # 3xN
over = xyz > 0.2068966
xyz[over] = xyz[over] ** 3.
xyz[~over] = (xyz[~over] - 0.13793103448275862) / 7.787
# Convert XYZ->LAB
rgbs = np.dot(_xyz2rgb_norm, xyz).T
over = rgbs > 0.0031308
rgbs[over] = 1.055 * (rgbs[over] ** (1. / 2.4)) - 0.055
rgbs[~over] *= 12.92
if n_dim == 4:
rgbs = np.concatenate((rgbs, labs[:, 3]), axis=1)
rgbs = np.clip(rgbs, 0., 1.)
return rgbs
|
