CODEWITHSUNDEEP
python-3.xnumpy
YeongHwa Jin
YeongHwa Jin

Reputation: 455

How can I make 1,300,000 images into one array file?

I have about 1,300,000 images. To reduce I/O cost, I want to make one file out of all of the images. So, file need to be (1300000, 256, 256, 3).

I intended to solve it with numpy, unfortunately, I can't read and write all the images at once because I have just 16gb ram.

Could you suggest some way to solve it efficiently?

Upvotes: 0

Views: 40

Answers (1)

Colwin
Colwin

Reputation: 2685

You will need to concatenate the images one at a time. Here is a possible solution using PIL/Pillow

import sys
import glob
from PIL import Image

def concat_two_images(image1, image2):
    images = map(Image.open, [image1, image2])
    widths, heights = zip(*(i.size for i in images))

    total_width = sum(widths)
    max_height = max(heights)

    new_im = Image.new('RGB', (total_width, max_height))

    x_offset = 0
    for im in images:
        new_im.paste(im, (x_offset,0))
        x_offset += im.size[0]

    new_im.save('concat_image.jpg')
    return 'concat_image.jpg'


first_concat_image = concat_two_images('test_image1.jpg', 'test_image2.jpg')
for image in glob.glob('*.jpg'):
     # TODO - Add code to skip the first 2 images
     new_image = concat_two_images(first_concat_image, image)

Upvotes: 1

Related Questions