从 Julia 文件调用函数不会读取 Python 文件中的图像

Question

In Julia I have built a function that changes an image a little bit. The function takes in as input following type

Matrix{RGB{N0f8}} (alias for Array{RGB{Normed{UInt8, 8}}, 2})

I have created environment where I use Julia and Python. So in Python I want to test function on the cifar database. I want to change the images in the cifar database with my function. But whenever I try to run the function from Julia on images in the Python file, I get this error:

JULIA: BoundsError: attempt to access 100×100×3 Array{Float32, 3} at index [1, 1]

Even through if I try to access this coordinates directly like this in Python:

image[1, 1]

it is completely fine and gives me the rgb values. So I was curious maybe there is some function from images that I could apply?

Answer 1

100×100×3 Array{Float32, 3} at index [1, 1]

you need 3 numbers to index a single element of this 3D array. If I have to guess, I would think the R pixel of the image[1, 1] in Python would be image[1, 1, 1] in Julia, and the G pixel is image[1, 1, 2] , and the B pixel is image[1, 1, 3]

---

looking at this example: https://juliaimages.org/latest/tutorials/arrays_colors/#fixedpoint

I think given a 3D array, you can convert it to an RGB matrix by using this from Images.jl :

using Images

img = colorview(RGB, reshape(image, (3, 100, 100)))