Getting the standard deviation of an image in Julia - image

Question

I'm using the Images package and I want to load in an image and get the mean and standard deviation of the pixels in r image.
I tried:
using Images, Color, FixedPointNumbers, ImageView, Testimages
img = testimage("mandril")
mean(data(img))
The mean worked fine, in fact in IJulia it displays the mean color of the image. However, when I tried to get the standard deviation of the image, I get:
std(data(img))
`varm` has no method matching varm(::Image{RGB{UfixedBase{Uint8,8}},2,Array{RGB{UfixedBase{Uint8,8}},2}}, ::RGB{Float32})
while loading In[66], in expression starting on line 1
in var at statistics.jl:162
How would one go about getting the standard deviation of the image?

Answer 1

You can use red, green and blue to extract the components of image,
and compute the standard deviation with std on the resulting matrices.
using Images
using TestImages
using Color
img = testimage("mandril")
data(img)
RGB(
std(red(img)),
std(green(img)),
std(blue(img))
)
# RGB{Float32}(0.22030124f0,0.18964756f0,0.24422659f0)
You could also build a 3-dimensional array with all the data,
with separate,
and apply std on the three 2-dimensional slices corresponding
to the three components, with mapslices.
vec( mapslices( std, separate(img), [1,2] ) )
If you want to figure out why mean(data(img)) works
and std(data(img)) does not (at least, currently),
you can check which method is called
with #which, and read the corresponding code, with #less.
#which mean(data(img))
# mean(A::AbstractArray{T,N}) at statistics.jl:17
#which std(data(img))
# std(A::AbstractArray{T,N}) at statistics.jl:204
#less mean(data(img))
#less std(data(img))
There are (at least) two problems: first, sqrt is not defined for colours,
second, some of the code in std assumes that the mean is a Number.