Coverage for difference/delta

1"""

2:math:`\\Delta E^*_{ab}` - Delta E Colour Difference

3====================================================

5Define the :math:`\\Delta E^*_{ab}` colour difference computation objects.

7- :attr:`colour.difference.JND_CIE1976`

8- :func:`colour.difference.delta_E_CIE1976`

9- :func:`colour.difference.delta_E_CIE1994`

10- :func:`colour.difference.delta_E_CIE2000`

11- :func:`colour.difference.delta_E_CMC`

12- :func:`colour.difference.delta_E_ITP`

13- :func:`colour.difference.delta_E_HyAB`

14- :func:`colour.difference.delta_E_HyCH`

16References

17----------

18- :cite:`Abasi2020a` : Abasi, S., Amani Tehran, M., & Fairchild, M. D. (2020).

19 Distance metrics for very large color differences. Color Research &

20 Application, 45(2), 208-223. doi:10.1002/col.22451

21- :cite:`InternationalTelecommunicationUnion2019` : International

22 Telecommunication Union. (2019). Recommendation ITU-R BT.2124-0 -

23 Objective metric for the assessment of the potential visibility of colour

24 differences in television (pp. 1-36).

25 https://www.itu.int/dms_pubrec/itu-r/rec/bt/R-REC-BT.2124-0-201901-I!!PDF-E.pdf

26- :cite:`Lindbloom2003c` : Lindbloom, B. (2003). Delta E (CIE 1976).

27 Retrieved February 24, 2014, from

28 http://brucelindbloom.com/Eqn_DeltaE_CIE76.html

29- :cite:`Lindbloom2009f` : Lindbloom, B. (2009). Delta E (CMC). Retrieved

30 February 24, 2014, from http://brucelindbloom.com/Eqn_DeltaE_CMC.html

31- :cite:`Lindbloom2011a` : Lindbloom, B. (2011). Delta E (CIE 1994).

32 Retrieved February 24, 2014, from

33 http://brucelindbloom.com/Eqn_DeltaE_CIE94.html

34- :cite:`Melgosa2013b` : Melgosa, M. (2013). CIE / ISO new standard:

35 CIEDE2000. http://www.color.org/events/colorimetry/\

36Melgosa_CIEDE2000_Workshop-July4.pdf

37- :cite:`Sharma2005b` : Sharma, G., Wu, W., & Dalal, E. N. (2005). The

38 CIEDE2000 color-difference formula: Implementation notes, supplementary

39 test data, and mathematical observations. Color Research & Application,

40 30(1), 21-30. doi:10.1002/col.20070

41- :cite:`Mokrzycki2011` : Mokrzycki, W., & Tatol, M. (2011). Color difference

42 Delta E - A survey. Machine Graphics and Vision, 20, 383-411.

43"""

45from __future__ import annotations

47from dataclasses import astuple, dataclass, field

49import numpy as np

51from colour.algebra import euclidean_distance

52from colour.hints import ( # noqa: TC001

53 Domain1,

54 Domain100,

55 NDArrayFloat,

56)

57from colour.utilities import (

58 MixinDataclassArithmetic,

59 as_float,

60 to_domain_100,

61 tsplit,

62 zeros,

63)

64from colour.utilities.documentation import DocstringFloat, is_documentation_building

66__author__ = "Colour Developers"

68__license__ = "BSD-3-Clause - https://opensource.org/licenses/BSD-3-Clause"

69__maintainer__ = "Colour Developers"

70__email__ = "colour-developers@colour-science.org"

71__status__ = "Production"

73__all__ = [

74 "JND_CIE1976",

75 "delta_E_CIE1976",

76 "delta_E_CIE1994",

77 "intermediate_attributes_CIE2000",

78 "delta_E_CIE2000",

79 "delta_E_CMC",

80 "delta_E_ITP",

81 "delta_E_HyAB",

82 "delta_E_HyCH",

83]

85JND_CIE1976 = 2.3

86if is_documentation_building(): # pragma: no cover

87 JND_CIE1976 = DocstringFloat(JND_CIE1976)

88 JND_CIE1976.__doc__ = """

89Just Noticeable Difference (JND) according to *CIE 1976* colour difference

90formula, i.e., Euclidean distance in *CIE L\\*a\\*b\\** colourspace.

92Notes

93-----

94A standard observer sees the difference in colour as follows:

96- 0 < :math:`\\Delta E^*_{ab}` < 1 : Observer does not notice the difference.

97- 1 < :math:`\\Delta E^*_{ab}` < 2 : Only experienced observer can notice the

98 difference.

99- 2 < :math:`\\Delta E^*_{ab}` < 3:5 : Unexperienced observer also notices

100 the difference.

101- 3:5 < :math:`\\Delta E^*_{ab}` < 5 : Clear difference in colour is noticed.

102- 5 < :math:`\\Delta E^*_{ab}` : Observer notices two different colours.

103

104References

105----------

106:cite:`Mokrzycki2011`

107"""

108

109

110def delta_E_CIE1976(Lab_1: Domain100, Lab_2: Domain100) -> NDArrayFloat:

111 """

112 Compute the colour difference :math:`\\Delta E_{76}` between two

113 specified *CIE L\\*a\\*b\\** colourspace arrays using the *CIE 1976*

114 recommendation.

115

116 Parameters

117 ----------

118 Lab_1

119 *CIE L\\*a\\*b\\** colourspace array 1.

120 Lab_2

121 *CIE L\\*a\\*b\\** colourspace array 2.

122

123 Returns

124 -------

125 :class:`numpy.ndarray`

126 Colour difference :math:`\\Delta E_{76}`.

127

128 Notes

129 -----

130 +------------+-----------------------+-------------------+

131 | **Domain** | **Scale - Reference** | **Scale - 1** |

132 +============+=======================+===================+

133 | ``Lab_1`` | 100 | 1 |

134 +------------+-----------------------+-------------------+

135 | ``Lab_2`` | 100 | 1 |

136 +------------+-----------------------+-------------------+

137

138 References

139 ----------

140 :cite:`Lindbloom2003c`

141

142 Examples

143 --------

144 >>> Lab_1 = np.array([48.99183622, -0.10561667, 400.65619925])

145 >>> Lab_2 = np.array([50.65907324, -0.11671910, 402.82235718])

146 >>> delta_E_CIE1976(Lab_1, Lab_2) # doctest: +ELLIPSIS

147 2.7335037...

148 """

149

150 return euclidean_distance(to_domain_100(Lab_1), to_domain_100(Lab_2))

151

152

153def delta_E_CIE1994(

154 Lab_1: Domain100,

155 Lab_2: Domain100,

156 textiles: bool = False,

157) -> NDArrayFloat:

158 """

159 Compute the colour difference :math:`\\Delta E_{94}` between two specified

160 *CIE L\\*a\\*b\\** colourspace arrays using the *CIE 1994* recommendation.

161

162 Parameters

163 ----------

164 Lab_1

165 *CIE L\\*a\\*b\\** colourspace array 1.

166 Lab_2

167 *CIE L\\*a\\*b\\** colourspace array 2.

168 textiles

169 Textiles application specific parametric factors,

170 :math:`k_L=2,\\ k_C=k_H=1,\\ k_1=0.048,\\ k_2=0.014` weights are used

171 instead of :math:`k_L=k_C=k_H=1,\\ k_1=0.045,\\ k_2=0.015`.

172

173 Returns

174 -------

175 :class:`numpy.ndarray`

176 Colour difference :math:`\\Delta E_{94}`.

177

178 Notes

179 -----

180 +------------+-----------------------+-------------------+

181 | **Domain** | **Scale - Reference** | **Scale - 1** |

182 +============+=======================+===================+

183 | ``Lab_1`` | 100 | 1 |

184 +------------+-----------------------+-------------------+

185 | ``Lab_2`` | 100 | 1 |

186 +------------+-----------------------+-------------------+

187

188 - *CIE 1994* colour differences are not symmetrical: difference between

189 ``Lab_1`` and ``Lab_2`` may not be the same as difference between

190 ``Lab_2`` and ``Lab_1`` thus one colour must be understood to be the

191 reference against which a sample colour is compared.

192

193 References

194 ----------

195 :cite:`Lindbloom2011a`

196

197 Examples

198 --------

199 >>> Lab_1 = np.array([48.99183622, -0.10561667, 400.65619925])

200 >>> Lab_2 = np.array([50.65907324, -0.11671910, 402.82235718])

201 >>> delta_E_CIE1994(Lab_1, Lab_2) # doctest: +ELLIPSIS

202 1.6711191...

203 >>> delta_E_CIE1994(Lab_1, Lab_2, textiles=True) # doctest: +ELLIPSIS

204 0.8404677...

205 """

206

207 L_1, a_1, b_1 = tsplit(to_domain_100(Lab_1))

208 L_2, a_2, b_2 = tsplit(to_domain_100(Lab_2))

209

210 k_1 = 0.048 if textiles else 0.045

211 k_2 = 0.014 if textiles else 0.015

212 k_L = 2 if textiles else 1

213 k_C = 1

214 k_H = 1

215

216 C_1 = np.hypot(a_1, b_1)

217 C_2 = np.hypot(a_2, b_2)

218

219 s_L = 1

220 s_C = 1 + k_1 * C_1

221 s_H = 1 + k_2 * C_1

222

223 delta_L = L_1 - L_2

224 delta_C = C_1 - C_2

225 delta_A = a_1 - a_2

226 delta_B = b_1 - b_2

227

228 radical = delta_A**2 + delta_B**2 - delta_C**2

229 delta_H = zeros(radical.shape)

230 delta_H[radical > 0] = np.sqrt(radical[radical > 0])

231

232 L = (delta_L / (k_L * s_L)) ** 2

233 C = (delta_C / (k_C * s_C)) ** 2

234 H = (delta_H / (k_H * s_H)) ** 2

235

236 d_E = np.sqrt(L + C + H)

237

238 return as_float(d_E)

239

240

241@dataclass

242class Attributes_Specification_CIE2000(MixinDataclassArithmetic):

243 """

244 Define the *CIE 2000* colour-difference formula attribute specification.

245

246 Parameters

247 ----------

248 J

249 Correlate of *lightness* :math:`J`.

250 C

251 Correlate of *chroma* :math:`C`.

252 h

253 *Hue* angle :math:`h` in degrees.

254 s

255 Correlate of *saturation* :math:`s`.

256 Q

257 Correlate of *brightness* :math:`Q`.

258 M

259 Correlate of *colourfulness* :math:`M`.

260 H

261 *Hue* :math:`h` quadrature :math:`H`.

262 HC

263 *Hue* :math:`h` composition :math:`H^C`.

264 """

265

266 S_L: float | NDArrayFloat | None = field(default_factory=lambda: None)

267 S_C: float | NDArrayFloat | None = field(default_factory=lambda: None)

268 S_H: float | NDArrayFloat | None = field(default_factory=lambda: None)

269 delta_L_p: float | NDArrayFloat | None = field(default_factory=lambda: None)

270 delta_C_p: float | NDArrayFloat | None = field(default_factory=lambda: None)

271 delta_H_p: float | NDArrayFloat | None = field(default_factory=lambda: None)

272 R_T: float | NDArrayFloat | None = field(default_factory=lambda: None)

273

274

275def intermediate_attributes_CIE2000(

276 Lab_1: Domain100, Lab_2: Domain100

277) -> Attributes_Specification_CIE2000:

278 """

279 Compute intermediate attributes for CIE 2000 colour difference calculation

280 between two specified *CIE L\\*a\\*b\\** colourspace arrays.

281

282 The intermediate attributes include the lightness, chroma, and hue

283 weighting functions (S_L, S_C, S_H), as well as the adjusted colour

284 differences (delta_L_p, delta_C_p, delta_H_p) and the rotation term (R_T)

285 required for computing :math:`\\Delta E_{00}`.

286

287 Parameters

288 ----------

289 Lab_1

290 *CIE L\\*a\\*b\\** colourspace array 1.

291 Lab_2

292 *CIE L\\*a\\*b\\** colourspace array 2.

293

294 Returns

295 -------

296 :class:`numpy.ndarray`

297 Intermediate attributes to compute the difference :math:`\\Delta E_{00}`.

298

299 Notes

300 -----

301 +------------+-----------------------+-------------------+

302 | **Domain** | **Scale - Reference** | **Scale - 1** |

303 +============+=======================+===================+

304 | ``Lab_1`` | 100 | 1 |

305 +------------+-----------------------+-------------------+

306 | ``Lab_2`` | 100 | 1 |

307 +------------+-----------------------+-------------------+

308

309 References

310 ----------

311 :cite:`Melgosa2013b`, :cite:`Sharma2005b`

312

313 Examples

314 --------

315 >>> Lab_1 = np.array([48.99183622, -0.10561667, 400.65619925])

316 >>> Lab_2 = np.array([50.65907324, -0.11671910, 402.82235718])

317 >>> intermediate_attributes_CIE2000(Lab_1, Lab_2) # doctest: +ELLIPSIS

318 Attributes_Specification_CIE2000(S_L=1.0001021..., S_C=19.0782682..., \

319S_H=4.7226695..., delta_L_p=1.6672370..., delta_C_p=2.1661609..., \

320delta_H_p=0.0105030..., R_T=-3...)

321 """

322

323 L_1, a_1, b_1 = tsplit(to_domain_100(Lab_1))

324 L_2, a_2, b_2 = tsplit(to_domain_100(Lab_2))

325

326 C_1_ab = np.hypot(a_1, b_1)

327 C_2_ab = np.hypot(a_2, b_2)

328

329 C_bar_ab = (C_1_ab + C_2_ab) / 2

330 C_bar_ab_7 = C_bar_ab**7

331

332 G = 0.5 * (1 - np.sqrt(C_bar_ab_7 / (C_bar_ab_7 + 25**7)))

333

334 a_p_1 = (1 + G) * a_1

335 a_p_2 = (1 + G) * a_2

336

337 C_p_1 = np.hypot(a_p_1, b_1)

338 C_p_2 = np.hypot(a_p_2, b_2)

339

340 h_p_1 = np.where(

341 np.logical_and(b_1 == 0, a_p_1 == 0),

342 0,

343 np.degrees(np.arctan2(b_1, a_p_1)) % 360,

344 )

345 h_p_2 = np.where(

346 np.logical_and(b_2 == 0, a_p_2 == 0),

347 0,

348 np.degrees(np.arctan2(b_2, a_p_2)) % 360,

349 )

350

351 delta_L_p = L_2 - L_1

352

353 delta_C_p = C_p_2 - C_p_1

354

355 h_p_2_s_1 = h_p_2 - h_p_1

356 C_p_1_m_2 = C_p_1 * C_p_2

357 delta_h_p = np.select(

358 [

359 C_p_1_m_2 == 0,

360 np.fabs(h_p_2_s_1) <= 180,

361 h_p_2_s_1 > 180,

362 h_p_2_s_1 < -180,

363 ],

364 [

365 0,

366 h_p_2_s_1,

367 h_p_2_s_1 - 360,

368 h_p_2_s_1 + 360,

369 ],

370 )

371

372 delta_H_p = 2 * np.sqrt(C_p_1_m_2) * np.sin(np.deg2rad(delta_h_p / 2))

373

374 L_bar_p = (L_1 + L_2) / 2

375

376 C_bar_p = (C_p_1 + C_p_2) / 2

377

378 a_h_p_1_s_2 = np.fabs(h_p_1 - h_p_2)

379 h_p_1_a_2 = h_p_1 + h_p_2

380 h_bar_p = np.select(

381 [

382 C_p_1_m_2 == 0,

383 a_h_p_1_s_2 <= 180,

384 np.logical_and(a_h_p_1_s_2 > 180, h_p_1_a_2 < 360),

385 np.logical_and(a_h_p_1_s_2 > 180, h_p_1_a_2 >= 360),

386 ],

387 [

388 h_p_1_a_2,

389 h_p_1_a_2 / 2,

390 (h_p_1_a_2 + 360) / 2,

391 (h_p_1_a_2 - 360) / 2,

392 ],

393 )

394

395 T = (

396 1

397 - 0.17 * np.cos(np.deg2rad(h_bar_p - 30))

398 + 0.24 * np.cos(np.deg2rad(2 * h_bar_p))

399 + 0.32 * np.cos(np.deg2rad(3 * h_bar_p + 6))

400 - 0.20 * np.cos(np.deg2rad(4 * h_bar_p - 63))

401 )

402

403 delta_theta = 30 * np.exp(-(((h_bar_p - 275) / 25) ** 2))

404

405 C_bar_p_7 = C_bar_p**7

406 R_C = 2 * np.sqrt(C_bar_p_7 / (C_bar_p_7 + 25**7))

407

408 L_bar_p_2 = (L_bar_p - 50) ** 2

409 S_L = 1 + ((0.015 * L_bar_p_2) / np.sqrt(20 + L_bar_p_2))

410

411 S_C = 1 + 0.045 * C_bar_p

412

413 S_H = 1 + 0.015 * C_bar_p * T

414

415 R_T = -np.sin(np.deg2rad(2 * delta_theta)) * R_C

416

417 return Attributes_Specification_CIE2000(

418 S_L,

419 S_C,

420 S_H,

421 delta_L_p,

422 delta_C_p,

423 delta_H_p,

424 R_T,

425 )

426

427

428def delta_E_CIE2000(

429 Lab_1: Domain100,

430 Lab_2: Domain100,

431 textiles: bool = False,

432) -> NDArrayFloat:

433 """

434 Compute the colour difference :math:`\\Delta E_{00}` between two specified

435 *CIE L\\*a\\*b\\** colourspace arrays using the *CIE 2000* recommendation.

436

437 Parameters

438 ----------

439 Lab_1

440 *CIE L\\*a\\*b\\** colourspace array 1.

441 Lab_2

442 *CIE L\\*a\\*b\\** colourspace array 2.

443 textiles

444 Textiles application specific parametric factors.

445 :math:`k_L=2,\\ k_C=k_H=1` weights are used instead of

446 :math:`k_L=k_C=k_H=1`.

447

448 Returns

449 -------

450 :class:`numpy.ndarray`

451 Colour difference :math:`\\Delta E_{00}`.

452

453 Notes

454 -----

455 +------------+-----------------------+-------------------+

456 | **Domain** | **Scale - Reference** | **Scale - 1** |

457 +============+=======================+===================+

458 | ``Lab_1`` | 100 | 1 |

459 +------------+-----------------------+-------------------+

460 | ``Lab_2`` | 100 | 1 |

461 +------------+-----------------------+-------------------+

462

463 - Parametric factors :math:`k_L=k_C=k_H=1` weights under

464 *reference conditions*:

465

466 - Illumination: D65 source

467 - Illuminance: 1000 lx

468 - Observer: Normal colour vision

469 - Background field: Uniform, neutral gray with :math:`L^*=50`

470 - Viewing mode: Object

471 - Sample size: Greater than 4 degrees

472 - Sample separation: Direct edge contact

473 - Sample colour-difference magnitude: Lower than 5.0

474 :math:`\\Delta E_{00}`

475 - Sample structure: Homogeneous (without texture)

476

477 References

478 ----------

479 :cite:`Melgosa2013b`, :cite:`Sharma2005b`

480

481 Examples

482 --------

483 >>> Lab_1 = np.array([48.99183622, -0.10561667, 400.65619925])

484 >>> Lab_2 = np.array([50.65907324, -0.11671910, 402.82235718])

485 >>> delta_E_CIE2000(Lab_1, Lab_2) # doctest: +ELLIPSIS

486 1.6709303...

487 >>> delta_E_CIE2000(Lab_1, Lab_2, textiles=True) # doctest: +ELLIPSIS

488 0.8412338...

489 """

490

491 S_L, S_C, S_H, delta_L_p, delta_C_p, delta_H_p, R_T = astuple(

492 intermediate_attributes_CIE2000(Lab_1, Lab_2)

493 )

494

495 k_L = 2 if textiles else 1

496 k_C = 1

497 k_H = 1

498

499 d_E = np.sqrt(

500 (delta_L_p / (k_L * S_L)) ** 2

501 + (delta_C_p / (k_C * S_C)) ** 2

502 + (delta_H_p / (k_H * S_H)) ** 2

503 + R_T * (delta_C_p / (k_C * S_C)) * (delta_H_p / (k_H * S_H))

504 )

505

506 return as_float(d_E)

507

508

509def delta_E_CMC(

510 Lab_1: Domain100,

511 Lab_2: Domain100,

512 l: float = 2, # noqa: E741

513 c: float = 1,

514) -> NDArrayFloat:

515 """

516 Compute the colour difference :math:`\\Delta E_{CMC}` between two

517 specified *CIE L\\*a\\*b\\** colourspace arrays using the *Colour

518 Measurement Committee* recommendation.

519

520 The quasimetric has two parameters: *lightness* (l) and *chroma* (c),

521 allowing users to weight the difference based on the ratio of l:c.

522 Commonly used values are 2:1 for acceptability and 1:1 for the

523 threshold of imperceptibility.

524

525 Parameters

526 ----------

527 Lab_1

528 *CIE L\\*a\\*b\\** colourspace array 1.

529 Lab_2

530 *CIE L\\*a\\*b\\** colourspace array 2.

531 l

532 *Lightness* weighting factor.

533 c

534 *Chroma* weighting factor.

535

536 Returns

537 -------

538 :class:`numpy.ndarray`

539 Colour difference :math:`\\Delta E_{CMC}`.

540

541 Notes

542 -----

543 +------------+-----------------------+-------------------+

544 | **Domain** | **Scale - Reference** | **Scale - 1** |

545 +============+=======================+===================+

546 | ``Lab_1`` | 100 | 1 |

547 +------------+-----------------------+-------------------+

548 | ``Lab_2`` | 100 | 1 |

549 +------------+-----------------------+-------------------+

550

551 References

552 ----------

553 :cite:`Lindbloom2009f`

554

555 Examples

556 --------

557 >>> Lab_1 = np.array([48.99183622, -0.10561667, 400.65619925])

558 >>> Lab_2 = np.array([50.65907324, -0.11671910, 402.82235718])

559 >>> delta_E_CMC(Lab_1, Lab_2) # doctest: +ELLIPSIS

560 0.8996999...

561 """

562

563 L_1, a_1, b_1 = tsplit(to_domain_100(Lab_1))

564 L_2, a_2, b_2 = tsplit(to_domain_100(Lab_2))

565

566 C_1 = np.hypot(a_1, b_1)

567 C_2 = np.hypot(a_2, b_2)

568 s_L = np.where(L_1 < 16, 0.511, (0.040975 * L_1) / (1 + 0.01765 * L_1))

569 s_C = 0.0638 * C_1 / (1 + 0.0131 * C_1) + 0.638

570 h_1 = np.degrees(np.arctan2(b_1, a_1)) % 360

571

572 t = np.where(

573 np.logical_and(h_1 >= 164, h_1 <= 345),

574 0.56 + np.fabs(0.2 * np.cos(np.deg2rad(h_1 + 168))),

575 0.36 + np.fabs(0.4 * np.cos(np.deg2rad(h_1 + 35))),

576 )

577

578 C_4 = C_1 * C_1 * C_1 * C_1

579 f = np.sqrt(C_4 / (C_4 + 1900))

580 s_h = s_C * (f * t + 1 - f)

581

582 delta_L = L_1 - L_2

583 delta_C = C_1 - C_2

584 delta_A = a_1 - a_2

585 delta_B = b_1 - b_2

586 delta_H2 = delta_A**2 + delta_B**2 - delta_C**2

587

588 v_1 = delta_L / (l * s_L)

589 v_2 = delta_C / (c * s_C)

590 v_3 = s_h

591

592 d_E = np.sqrt(v_1**2 + v_2**2 + (delta_H2 / (v_3 * v_3)))

593

594 return as_float(d_E)

595

596

597def delta_E_ITP(ICtCp_1: Domain1, ICtCp_2: Domain1) -> NDArrayFloat:

598 """

599 Compute the colour difference :math:`\\Delta E_{ITP}` between two specified

600 :math:`IC_TC_P` colour encoding arrays using the

601 *Recommendation ITU-R BT.2124*.

602

603 Parameters

604 ----------

605 ICtCp_1

606 :math:`IC_TC_P` colour encoding array 1.

607 ICtCp_2

608 :math:`IC_TC_P` colour encoding array 2.

609

610 Returns

611 -------

612 :class:`numpy.ndarray`

613 Colour difference :math:`\\Delta E_{ITP}`.

614

615 Notes

616 -----

617 - A value of 1 is equivalent to a just noticeable difference when viewed

618 in the most critical adaptation state.

619

620 +--------------+-----------------------+--------------------+

621 | **Domain** | **Scale - Reference** | **Scale - 1** |

622 +==============+=======================+====================+

623 | ``ICtCp_1`` | 1 | 1 |

624 +--------------+-----------------------+--------------------+

625 | ``ICtCp_2`` | 1 | 1 |

626 +--------------+-----------------------+--------------------+

627

628 References

629 ----------

630 :cite:`InternationalTelecommunicationUnion2019`

631

632 Examples

633 --------

634 >>> ICtCp_1 = np.array([0.4885468072, -0.04739350675, 0.07475401302])

635 >>> ICtCp_2 = np.array([0.4899203231, -0.04567508203, 0.07361341775])

636 >>> delta_E_ITP(ICtCp_1, ICtCp_2) # doctest: +ELLIPSIS

637 1.42657228...

638 """

639

640 I_1, T_1, P_1 = tsplit(ICtCp_1)

641 T_1 *= 0.5

642

643 I_2, T_2, P_2 = tsplit(ICtCp_2)

644 T_2 *= 0.5

645

646 d_E_ITP = 720 * np.sqrt(

647 ((I_2 - I_1) ** 2) + ((T_2 - T_1) ** 2) + ((P_2 - P_1) ** 2)

648 )

649

650 return as_float(d_E_ITP)

651

652

653def delta_E_HyAB(Lab_1: Domain100, Lab_2: Domain100) -> NDArrayFloat:

654 """

655 Compute the colour difference between two *CIE L\\*a\\*b\\** colourspace arrays

656 using a combination of a Euclidean metric in hue and chroma with a

657 city-block metric to incorporate lightness differences.

658

659 This metric is intended for large colour differences, on the order of 10

660 *CIE L\\*a\\*b\\** units or greater.

661

662 Parameters

663 ----------

664 Lab_1

665 *CIE L\\*a\\*b\\** colourspace array 1.

666 Lab_2

667 *CIE L\\*a\\*b\\** colourspace array 2.

668

669 Returns

670 -------

671 :class:`numpy.ndarray`

672 Colour difference :math:`\\Delta E_{HyAB}`.

673

674 Notes

675 -----

676 +------------+-----------------------+-------------------+

677 | **Domain** | **Scale - Reference** | **Scale - 1** |

678 +============+=======================+===================+

679 | ``Lab_1`` | 100 | 1 |

680 +------------+-----------------------+-------------------+

681 | ``Lab_2`` | 100 | 1 |

682 +------------+-----------------------+-------------------+

683

684 References

685 ----------

686 :cite:`Abasi2020a`

687

688 Examples

689 --------

690 >>> Lab_1 = np.array([39.91531343, 51.16658481, 146.12933781])

691 >>> Lab_2 = np.array([53.12207516, -39.92365056, 249.54831278])

692 >>> delta_E_HyAB(Lab_1, Lab_2) # doctest: +ELLIPSIS

693 151.0215481...

694 """

695

696 dLab = to_domain_100(Lab_1) - to_domain_100(Lab_2)

697 dL, da, db = tsplit(dLab)

698 HyAB = np.abs(dL) + np.hypot(da, db)

699

700 return as_float(HyAB)

701

702

703def delta_E_HyCH(

704 Lab_1: Domain100,

705 Lab_2: Domain100,

706 textiles: bool = False,

707) -> NDArrayFloat:

708 """

709 Compute the colour difference between two *CIE L\\*a\\*b\\** colourspace

710 arrays using a combination of Euclidean metric in hue and chroma with a

711 city-block metric to incorporate lightness differences based on

712 *CIE 2000* recommendation attributes.

713

714 This metric is intended for large colour differences, on the order of 10

715 *CIE L\\*a\\*b\\** units or greater.

716

717 Parameters

718 ----------

719 Lab_1

720 *CIE L\\*a\\*b\\** colourspace array 1.

721 Lab_2

722 *CIE L\\*a\\*b\\** colourspace array 2.

723 textiles

724 Whether to use the textile-specific parametrization.

725

726 Returns

727 -------

728 :class:`numpy.ndarray`

729 Colour difference :math:`\\Delta E_{HyCH}`.

730

731 Notes

732 -----

733 +------------+-----------------------+-------------------+

734 | **Domain** | **Scale - Reference** | **Scale - 1** |

735 +============+=======================+===================+

736 | ``Lab_1`` | 100 | 1 |

737 +------------+-----------------------+-------------------+

738 | ``Lab_2`` | 100 | 1 |

739 +------------+-----------------------+-------------------+

740

741 References

742 ----------

743 :cite:`Abasi2020a`

744

745 Examples

746 --------

747 >>> Lab_1 = np.array([39.91531343, 51.16658481, 146.12933781])

748 >>> Lab_2 = np.array([53.12207516, -39.92365056, 249.54831278])

749 >>> delta_E_HyCH(Lab_1, Lab_2) # doctest: +ELLIPSIS

750 48.664279419760369...

751 """

752

753 S_L, S_C, S_H, delta_L_p, delta_C_p, delta_H_p, R_T = astuple(

754 intermediate_attributes_CIE2000(Lab_1, Lab_2)

755 )

756

757 k_L = 2 if textiles else 1

758 k_C = 1

759 k_H = 1

760

761 HyCH = np.abs(delta_L_p / (k_L * S_L)) + np.sqrt(

762 (delta_C_p / (k_C * S_C)) ** 2 + (delta_H_p / (k_H * S_H)) ** 2

763 )

764

765 return as_float(HyCH)

Coverage for difference/delta_e.py: 76%

134 statements