Skip to content

Commit

Permalink
Use Float32 to calculate colordiff for low precision colors (#504)
Browse files Browse the repository at this point in the history 
For consistency with `DE_2000`, this allows `Float32` to be used for other metrics.
  • Loading branch information
@kimikage
kimikage committed Jul 30, 2021
1 parent 15466d6 commit f71d7a1
Showing 1 changed file with 79 additions and 68 deletions.
147 changes: 79 additions & 68 deletions src/differences.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -272,46 +272,49 @@ const DE2000_SINEXP_F32 = [Float32(π/3 * exp(-i)) for i = 0.0:0.25:87.25]
end

# Delta E94
function _colordiff(ai::Color, bi::Color, m::DE_94)
a = convert(LCHab, ai)
b = convert(LCHab, bi)
function _colordiff(a::Color, b::Color, m::DE_94)
_colordiff(promote(convert(Lab, a), convert(Lab, b))..., m)
end
function _colordiff(a::C, b::C, m::DE_94) where {T, C <: Union{Lab{T}, LCHab{T}}}
F = typeof(0.5f0 * zero(T)) === Float32 ? Float32 : promote_type(Float64, T)

# Calculate the delta values for each channel
dl, dc, dh = b.l - a.l, b.c - a.c, delta_h(b, a)
dl, dc, dh = b.l - a.l, delta_c(b, a), delta_h(b, a)

# Lightness, hue, chroma correction terms
sl = 1
sc = 1 + m.k1 * a.c
sh = 1 + m.k2 * a.c
# sl = 1
sc = muladd(F(m.k1), chroma(a), 1)
sh = muladd(F(m.k2), chroma(a), 1)

sqrt((dl/(m.kl*sl))^2 + (dc/(m.kc*sc))^2 + (dh/(m.kh*sh))^2)
sqrt((dl/F(m.kl))^2 + (dc/(F(m.kc)*sc))^2 + (dh/(F(m.kh)*sh))^2)
end

# Metric form of jpc79 color difference equation (mostly obsolete)
function _colordiff(ai::Color, bi::Color, m::DE_JPC79)
# Convert directly into LCh
a = convert(LCHab, ai)
b = convert(LCHab, bi)
function _colordiff(a::Color, b::Color, m::DE_JPC79)
_colordiff(promote(convert(Lab, a), convert(Lab, b))..., m)
end
function _colordiff(a::C, b::C, ::DE_JPC79) where {T, C <: Union{Lab{T}, LCHab{T}}}
F = typeof(0.5f0 * zero(T)) === Float32 ? Float32 : promote_type(Float64, T)

# Calculate deltas in each direction
dl, dc, dh = b.l - a.l, b.c - a.c, delta_h(b, a)
dl, dc, dh = b.l - a.l, delta_c(b, a), delta_h(b, a)

# Calculate mean lightness, chroma and hue
ml, mc, mh = (a.l + b.l) / 2, (a.c + b.c) / 2, mean_hue(a, b)
ml, mc, mh = (a.l + b.l) * F(0.5), (chroma(a) + chroma(b)) * F(0.5), mean_hue(a, b)

# L* adjustment term
sl = 0.08195*ml/(1+0.01765*ml)
sl = F(0.08195) * ml / muladd(F(0.01765), ml, 1)

# C* adjustment term
sc = 0.638+(0.0638*mc/(1+0.0131*mc))
sc = F(0.638) + F(0.0638) * mc / muladd(F(0.0131), mc, 1)

# H* adjustment term
if (mc < 0.38)
if mc < F(0.38)
sh = sc
elseif (mh >= 164 && mh <= 345)
sh = sc*(0.56+abs(0.2*cosd(mh+168)))
elseif mh >= 164 && mh <= 345
sh = sc * muladd(F(0.2), abs(cosd(mh + 168)), F(0.56))
else
sh = sc*(0.38+abs(0.4*cosd(mh+35)))
sh = sc * muladd(F(0.4), abs(cosd(mh + 35)), F(0.38))
end

# Calculate the final difference
Expand All @@ -320,94 +323,102 @@ end


# Metric form of the cmc color difference
function _colordiff(ai::Color, bi::Color, m::DE_CMC)
# Convert directly into LCh
a = convert(LCHab, ai)
b = convert(LCHab, bi)
function _colordiff(a::Color, b::Color, m::DE_CMC)
_colordiff(promote(convert(Lab, a), convert(Lab, b))..., m)
end
function _colordiff(a::C, b::C, m::DE_CMC) where {T, C <: Union{Lab{T}, LCHab{T}}}
F = typeof(0.5f0 * zero(T)) === Float32 ? Float32 : promote_type(Float64, T)

ac, ah = chroma(a), hue(a)

# Calculate deltas in each direction
dl, dc, dh = b.l - a.l, b.c - a.c, delta_h(b, a)
dl, dc, dh = b.l - a.l, delta_c(b, a), delta_h(b, a)

# L* adjustment term
if (a.l <= 16)
sl = 0.511
if a.l < 16
sl = F(0.511)
else
sl = 0.040975 * a.l / (1 + 0.01765 * a.l)
sl = F(0.040975) * a.l / muladd(F(0.01765), a.l, 1)
end

# C* adjustment term
sc = 0.0638 * a.c / (1 + 0.0131 * a.c) + 0.638
sc = F(0.0638) * ac / muladd(F(0.0131), ac, 1) + F(0.638)

f = sqrt(a.c^4 / (a.c^4 + 1900))
f = sqrt(ac^4 / (ac^4 + 1900))

if 164 <= a.h <= 345
t = 0.56 + abs(0.2 * cosd(a.h + 168))
if 164 <= ah <= 345
t = muladd(F(0.2), abs(cosd(ah + 168)), F(0.56))
else
t = 0.36 + abs(0.4 * cosd(a.h + 35))
t = muladd(F(0.4), abs(cosd(ah + 35)), F(0.36))
end

# H* adjustment term
sh = sc*(t*f+1-f)
sh = sc * muladd(t, f, 1 - f)

# Calculate the final difference
sqrt((dl/(m.kl*sl))^2 + (dc/(m.kc*sc))^2 + (dh/sh)^2)
sqrt((dl/(F(m.kl)*sl))^2 + (dc/(F(m.kc)*sc))^2 + (dh/sh)^2)
end

# The BFD color difference equation
function _colordiff(ai::Color, bi::Color, m::DE_BFD)
function _colordiff(a::Color, b::Color, m::DE_BFD)
# We have to start back in XYZ because BFD uses a different L equation
# Currently, support for the `wp` argument of `convert` is limited.
function to_xyz(c::Color, wp)
function to_xyz(c::Color{T}, wp) where T
F = typeof(0.5f0 * zero(T)) === Float32 ? Float32 : promote_type(Float64, T)
c isa XYZ && return c
(c isa xyY || c isa LMS) && return convert(XYZ, c)
(c isa Lab || c isa Luv) && return convert(XYZ, c, wp)
c isa LCHuv && return convert(XYZ, convert(Luv, c), wp)
convert(XYZ, convert(Lab, c), wp)
wpf = XYZ{F}(wp)
(c isa Lab || c isa Luv) && return convert(XYZ, c, wpf)
c isa LCHuv && return convert(XYZ, convert(Luv, c), wpf)
convert(XYZ, convert(Lab, c), wpf)
end
_colordiff(promote(to_xyz(a, m.wp), to_xyz(b, m.wp))..., m)
end
function _colordiff(a_xyz::XYZ{T}, b_xyz::XYZ{T}, m::DE_BFD) where T
F = typeof(0.5f0 * zero(T)) === Float32 ? Float32 : promote_type(Float64, T)

# We have to start back in XYZ because BFD uses a different L equation
a_XYZ = to_xyz(ai, m.wp)
b_XYZ = to_xyz(bi, m.wp)

la = 54.6*log10(a_XYZ.y+1.5)-9.6
lb = 54.6*log10(b_XYZ.y+1.5)-9.6
la = muladd(F(54.6), log10(a_xyz.y + F(1.5)), F(-9.6))
lb = muladd(F(54.6), log10(b_xyz.y + F(1.5)), F(-9.6))

# Convert into LCh with the proper white point
a_Lab = convert(Lab, a_XYZ, m.wp)
b_Lab = convert(Lab, b_XYZ, m.wp)
wpf = XYZ{F}(m.wp)
a_lab = convert(Lab, a_xyz, wpf)
b_lab = convert(Lab, b_xyz, wpf)

# Substitute in the different L values into the L*C*h values
a = LCHab(la, chroma(a_Lab), hue(a_Lab))
b = LCHab(lb, chroma(b_Lab), hue(b_Lab))
a = LCHab(la, chroma(a_lab), hue(a_lab))
b = LCHab(lb, chroma(b_lab), hue(b_lab))

# Calculate deltas in each direction
dl, dc, dh = b.l - a.l, b.c - a.c, delta_h(b, a)
dl, dc, dh = b.l - a.l, delta_c(b_lab, a_lab), delta_h(b_lab, a_lab)

# Find the mean value of the inputs to use as the "standard"
ml, mc, mh = (a.l + b.l) / 2, (a.c + b.c) / 2, mean_hue(a, b)
mc, mh = (a.c + b.c) * F(0.5), mean_hue(a, b)

# Correction terms for a variety of nonlinearities in CIELAB.
g = sqrt(mc^4/(mc^4 + 14000))
g = sqrt(mc^4 / (mc^4 + 14000))

t = 0.627 + 0.055 * cosd( mh - 245) -
0.040 * cosd(2mh - 136) +
0.070 * cosd(3mh - 32) +
0.049 * cosd(4mh + 114) -
0.015 * cosd(5mh + 103)
t_cos = F( 55) * cosd( mh - 245) +
F(-40) * cosd(2mh - 136) +
F( 70) * cosd(3mh - 32) +
F( 49) * cosd(4mh + 114) +
F(-15) * cosd(5mh + 103)
t = muladd(F(1/1000), t_cos, F(0.627))

rc = sqrt(mc^6/(mc^6 + 7e7))
rc = sqrt(mc^6 / (mc^6 + F(7e7)))

rh = -0.260 * cosd( mh - 308) -
0.379 * cosd(2mh - 160) -
0.636 * cosd(3mh - 254) +
0.226 * cosd(4mh + 140) -
0.194 * cosd(5mh + 280)
rh = F(-0.260) * cosd( mh - 308) +
F(-0.379) * cosd(2mh - 160) +
F(-0.636) * cosd(3mh - 254) +
F( 0.226) * cosd(4mh + 140) +
F(-0.194) * cosd(5mh + 280)

dcc = 0.035*mc/(1+0.00365*mc) + 0.521
dhh = dcc*(g*t+1-g)
rt = rc*rh
dcc = muladd(F(0.035), mc / muladd(F(0.00365), mc, 1) , F(0.521))
dhh = dcc * muladd(g, t, 1 - g)
rt = rc * rh

# Final calculation
sqrt((dl/m.kl)^2 + (dc/(m.kc*dcc))^2 + (dh/dhh)^2 + rt*((dc*dh)/(dcc*dhh)))
sqrt((dl/F(m.kl))^2 + (dc/(F(m.kc)*dcc))^2 + (dh/dhh)^2 + rt*((dc*dh)/(dcc*dhh)))
end

function _colordiff(ai::Color, bi::Color,
Expand Down

0 comments on commit f71d7a1

Please sign in to comment.