RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-25 16:27:35 +00:00
Code Issues Activity Actions

ICC: Add Profile::to_lab()

Browse source 
This can be used to convert a profile-dependent color to the L*a*b*
color space.

(I'd like to use this to implement the DeltaE (CIE 2000) algorithm,
which is a metric for how similar two colors are perceived.
(And I'd like to use that to evaluate color conversion roundtrip
quality, once I've implemented full conversions.)
This commit is contained in:
Nico Weber 2023-04-29 15:16:22 -04:00 • committed by Jelle Raaijmakers
parent 5fc11a316f
commit f3dbfb85d9
3 changed files with 110 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

50
Tests/LibGfx/TestICCProfile.cpp
View file

@ -153,3 +153,53 @@ TEST_CASE(to_pcs)
float f192 = sRGB_curve.evaluate(192 / 255.f);
EXPECT_APPROXIMATE_VECTOR3(xyz_from_sRGB(64, 128, 192), r_xyz * f64 + g_xyz * f128 + b_xyz * f192);
}
TEST_CASE(to_lab)
{
auto sRGB = MUST(Gfx::ICC::sRGB());
auto lab_from_sRGB = [sRGB](u8 r, u8 g, u8 b) {
u8 rgb[3] = { r, g, b };
return MUST(sRGB->to_lab(rgb));
};
// The `expected` numbers are from https://colorjs.io/notebook/ for this snippet of code:
// new Color("srgb", [0, 0, 0]).lab.toString();
//
// new Color("srgb", [1, 0, 0]).lab.toString();
// new Color("srgb", [0, 1, 0]).lab.toString();
// new Color("srgb", [0, 0, 1]).lab.toString();
//
// new Color("srgb", [1, 1, 0]).lab.toString();
// new Color("srgb", [1, 0, 1]).lab.toString();
// new Color("srgb", [0, 1, 1]).lab.toString();
//
// new Color("srgb", [1, 1, 1]).lab.toString();
Gfx::ICC::Profile::CIELAB expected[] = {
{ 0, 0, 0 },
{ 54.29054294696968, 80.80492033462421, 69.89098825896275 },
{ 87.81853633115202, -79.27108223854806, 80.99459785152247 },
{ 29.56829715344471, 68.28740665215547, -112.02971798617645 },
{ 97.60701009682253, -15.749846639252663, 93.39361164266089 },
{ 60.16894098715946, 93.53959546199253, -60.50080231921204 },
{ 90.66601315791455, -50.65651077286893, -14.961666625736525 },
{ 100.00000139649632, -0.000007807961277528364, 0.000006766250648659877 },
};
// We're off by more than the default EXPECT_APPROXIMATE() error, so use EXPECT_APPROXIMATE_WITH_ERROR().
// The difference is not too bad: ranges for L*, a*, b* are [0, 100], [-125, 125], [-125, 125],
// so this is an error of considerably less than 0.1 for u8 channels.
#define EXPECT_APPROXIMATE_LAB(l1, l2) \
EXPECT_APPROXIMATE_WITH_ERROR((l1).L, (l2).L, 0.01); \
EXPECT_APPROXIMATE_WITH_ERROR((l1).a, (l2).a, 0.03); \
EXPECT_APPROXIMATE_WITH_ERROR((l1).b, (l2).b, 0.02);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(0, 0, 0), expected[0]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(255, 0, 0), expected[1]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(0, 255, 0), expected[2]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(0, 0, 255), expected[3]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(255, 255, 0), expected[4]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(255, 0, 255), expected[5]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(0, 255, 255), expected[6]);
EXPECT_APPROXIMATE_LAB(lab_from_sRGB(255, 255, 255), expected[7]);
}