Invert structure of array

Question

I have an array representing a video. Currently the way it is set up is the first element is an array of all the values of the first pixel at all the different time steps, then the second element is of the second pixel and so on. I want it so the first element would be an array of all the pixels at the first time step.

So for two frames of a 2x2 video, I would want

[
  [
    [a1, a2, a3], [b1, b2, b3]
  ],
  [
    [c1, c2, c3], [d1, d2, d3]
  ]
]

to become

[
 [
  [a1, b1],
  [c1, d1]
 ],
 [
  [a2, b2],
  [c2, d2]
 ],
 [
  [a3, b3],
  [c3, d3]
 ],
]

My current implementation is this:

def remap_image(seq):
    # seq = np.array(seq)
    s = seq.shape
    a = np.zeros((s[2], s[0], s[1]))

    for x, px in enumerate(tqdm(seq)):
        for y, py in enumerate(px):
            for p_counter, value in enumerate(py):
                a[p_counter][x][y] = value/100.
    return a

This works as intended however this approach is incredibly slow. Is there any faster way to do this?

Looks like you have a (2,2,3) shape, and you want to transpose to (3,2,2) — hpaulj
– hpaulj, Commented Aug 28, 2021 at 5:22

Yu-Sheng Li · Accepted Answer · 2021-08-28 05:39:09Z

2

You can use einops with better semantics. For your case, the following simple code works.

from einops import rearrange

def remap_image(seq):
    return rearrange(seq, 'w h t -> t w h')

answered Aug 28, 2021 at 5:39

Yu-Sheng Li

946 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Mateen Ulhaq Over a year ago

Cool package. Apparently I've starred it but never used it after that. It is an additional dependency after all.

Boio · Accepted Answer · 2021-08-28 05:42:40Z

0

Following what Mateen Ulhaq said, simply

def remap_image(seq):
    return seq.transpose(2, 0, 1)

answered Aug 28, 2021 at 5:42

Boio

9271 gold badge7 silver badges15 bronze badges

Comments

Alain T. · Accepted Answer · 2021-08-28 06:10:57Z

0

I'm not sure it's going to be much faster but you could use zip():

a = [
  [
    ["a1", "a2", "a3"], ["b1", "b2", "b3"]
  ],
  [
    ["c1", "c2", "c3"], ["d1", "d2", "d3"]
  ]
]

b = (list(map(list,zip(*r))) for r in a)
b = list(map(list,zip(*b)))

print(b)

[
 [
    ['a1', 'b1'],
    ['c1', 'd1']
 ],
 [
    ['a2', 'b2'],
    ['c2', 'd2']
 ],
 [
    ['a3', 'b3'],
    ['c3', 'd3']
 ]
]

If your data is in a numpy array, then b = np.transpose(a,axes=(2,0,1)) would do it directly and very fast.

edited Aug 28, 2021 at 6:10

answered Aug 28, 2021 at 6:03

Alain T.

42.2k4 gold badges36 silver badges57 bronze badges

Collectives™ on Stack Overflow

Invert structure of array

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related