Proper way to make numpy array of ndarrays

Question

the following code achieves what I want to accomplish, but uses python lists and is probably very inefficient. Please let me know if there is a way to do the following purely with Numpy:

def makeImageArray(count):
    l = []
    for i in range (count):
        l.append(image)
    res = np.array(l)
    return res

Where image is a numpy array of shape (1200,1200,3).

Thank you so much!

You are doing the right thing - start with a list of images you want to join. — hpaulj
– hpaulj, Commented Mar 6, 2021 at 18:32

nckstr15 · Accepted Answer · 2021-03-06 18:29:10Z

1

It is possible using numpy.stack() (reference)

If you have multiple images you want to add to the new array you can use this

import numpy as np

image_0 = np.random.rand(1200,1200,3)
image_1 = np.random.rand(1200,1200,3)

stack = np.stack((image_0, image_1))
stack.shape

>>> (2, 1200, 1200, 3)

If you just want to stack one array multiple times

Edit

If you want to stack the same image:

image = np.random.rand(1200,1200,3)
count = 10
stack = np.stack([image for _ in range(count)])
stack.shape
>>> (10, 1200, 1200, 3)

edited Mar 6, 2021 at 18:29

answered Mar 6, 2021 at 18:23

nckstr15

1091 silver badge7 bronze badges

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

4 Comments

hpaulj Over a year ago

For this stack and array are the same.

nckstr15 Over a year ago

You are correct, but I think using stack makes the code more clear.

hpaulj Over a year ago

stack provides a different error/warning if the arrays don't match in size.

nckstr15 Over a year ago

Thats true, however it is deprecated/not good to append/store ndarrays with different shapes or lengths in an ndarray. And here it was onbly asked for images of the same size.

Collectives™ on Stack Overflow

Proper way to make numpy array of ndarrays

1 Answer 1

Edit

4 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Edit

4 Comments

Related