Create a 3D numpy array with a loop from multiple 2D images

Question

I want to read from many sample images, and make 2D numpy array that row "i" corresponds to i'th sample, and column j correspond to j'th pixel of image(all pixels of a 12*13 image is saved by a list of 156 numbers)

import numpy as np
images = np.array([])
for Letter in "ABCDEFGHIJKLMNO":
    im = Image.open("ABCDEFGHIJKLMNO\\16\\"+Letter+".bmp")
    sampleimage = list(im.getdata())
    images = np.append(images,[sampleimage])

however I struggling making 2D numpy array. above "images" array become an (1800,) array insted of (11,156) . I tried many differend ways but none of them working properly or efficient (making 2D python list and then converting to numpy array is not efficient. although even that solution doesn't work).

so my question is, what is the best way of creating a 2D numpy array on the fly?

np.append is hard to use correctly, and even then is slow. Reread the np.append docs. Without an axis parameter it flattens the inputs, hence your result. — hpaulj
– hpaulj, Commented Jul 5, 2018 at 16:41

Framester · Accepted Answer · 2018-07-05 16:07:36Z

1

You are treating the numpy array images like an list. Have a look at numpy.stack. But in my memory the fastest way is to convert each image to an numpy array, aggregate them in a list and then turn the list into an array.

import numpy as np
images = list()
for Letter in "ABCDEFGHIJKLMNO":
    im = Image.open("ABCDEFGHIJKLMNO\\16\\"+Letter+".bmp")
    sampleimage = np.array(im.getdata())
    images.append(sampleimage)
images_array = np.array(images)

In your case the resulting array should have the size [1800,11,156] or [11,156,1800]

answered Jul 5, 2018 at 16:07

Framester

35.9k52 gold badges137 silver badges194 bronze badges

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

Collectives™ on Stack Overflow

Create a 3D numpy array with a loop from multiple 2D images

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related