Trouble copying and reversing parts of array with numpy

Question

I am trying to copy a section of an input 2d array "img" and mirroring that section and copying it into the 2d array "out"

The following code does what I need

a = numpy.zeros(shape=(pad, pad))
a[:,:]=img[0:pad,0:pad]
out[0:pad,0:pad]=a[::-1,::-1]

But simply doing the following does not

out[0:pad,0:pad]=img[0:pad:-1,0:pad:-1]

and instead returns ValueError: could not broadcast input array from shape (0,0) into shape (2,2) for pad=2 and I am not sure why.

Answer 1

img[0:pad:-1,0:pad:-1] 

should be

img[pad-1::-1, pad-1::-1]

since you want the index to start at pad-1 and step down to 0. See here for the complete rules governing NumPy basic slicing .

For example,

import numpy as np

img = np.arange(24).reshape(6,4)
# array([[ 0,  1,  2,  3],
#        [ 4,  5,  6,  7],
#        [ 8,  9, 10, 11],
#        [12, 13, 14, 15],
#        [16, 17, 18, 19],
#        [20, 21, 22, 23]])

pad = 2
out = img[pad-1::-1, pad-1::-1]

print(out)

yields

[[5 4]
 [1 0]]