converting an Image into numpy array, then save into txt file, then again converting .txt in image

Question

I have three stages: 1. convert an Image into numpy array 2. save that array in a text file 3. convert the array into Image by reading that text file

I tried below code, which is converting image into an array and again reading the same array from the text file

from PIL import Image
import numpy
im = Image.open("a.jpg") # img size (480,910,3)

np_im = numpy.array(im)

with open('test.txt', 'w') as outfile:
    for slice_2d in np_im:
        numpy.savetxt(outfile, slice_2d)

new_data = numpy.loadtxt('test.txt')

new_data=new_data.reshape((480,910,3))

img = Image.fromarray(new_data,'RGB')
img.save('my.bmp')
img.show()

if i compare the array (before saving and after loading from file and reshaping) array looks exactly the same(except dot). ex.

[[[ 48  58  24]
  [ 48  58  24]
  [ 47  57  23]
... 
and 
[[[ 48.  58.  24.]
  [ 48.  58.  24.]
  [ 47.  57.  23.]
  ...

but the image I'm getting is completely distorted. why so?

Possible duplicate of [Image.fromarray just produces black image](https://stackoverflow.com/questions/47290668/image-fromarray-just-produces-black-image) — D Malan, Oct 31 '19 at 08:23

score 1 · Answer 1 · answered Oct 31 '19 at 08:32

1

after some exercise, I got my answer

img = Image.fromarray(new_data.astype(numpy.uint8),'RGB')

answered Oct 31 '19 at 08:32

sanghmitra

21
2
9

score 1 · Answer 2 · answered Oct 31 '19 at 08:40

You might be interested in numpy.savetxt. It sounds a lot like you're trying to build this.

There's also a complementary numpy.loadtxt.

If you feed savetxt a filename ending in .gz, it'll even compress the data. loadtxt understands it just the same.

This will work for arbitrary numpy arrays.

converting an Image into numpy array, then save into txt file, then again converting .txt in image

2 Answers2