Change the content of the array in python

Question

I have a multidimensional array(let's call it 'data'), I print it produce like this

[[[255, 255, 255]
  [255, 255, 255]
  [0, 0, 0]
  [255, 255, 255]]
 [[0, 0, 0]
  [255, 255, 255]
  [0, 0, 0]
  [0, 0, 0]]
  ... and so on
  [255, 255, 255]]]

i want to change the content of data like this

[[1,
  1,
  0,
  1]
 [0,
  1,
  0,
  0]
  ... and so on
  1]]

[255, 255, 255] become 1, and [0, 0, 0] become 0

I'm trying with numpy.where, but I'm desperate How to do that in python programming?

Are there going to be cases where the numbers are something else? Or are you strictly checking just between 3x255 and 3x0? — Muhammad Ali
– Muhammad Ali, Commented Apr 7, 2020 at 8:08
on your first code block, did you mean [255, 255, 255] or "[255 255 255]" — Aven Desta
– Aven Desta, Commented Apr 7, 2020 at 8:12
@baby idk for sure, but when printed, the result are like that — ircham
– ircham, Commented Apr 7, 2020 at 8:17

yatu · Accepted Answer · 2020-04-07 09:00:17Z

2

One way is checking whether a value is 255, and reducing the boolean result with np.logical_and

np.logical_and.reduce(a==255, axis=2).view('i1')

For the following example:

a = np.array([[[255, 255 ,255],
               [255, 255, 255],
               [0 ,0, 0],
               [255, 255 ,255]],
              [[0 ,0, 0],
               [255, 255 ,255],
               [0 ,0, 0],
               [0 ,0, 0]]])

np.logical_and.reduce(a==255, axis=2).view('i1') 

array([[1, 1, 0, 1],
       [0, 1, 0, 0]], dtype=int8)

edited Apr 7, 2020 at 9:00

answered Apr 7, 2020 at 8:08

yatu

88.6k12 gold badges93 silver badges148 bronze badges

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

3 Comments

ircham Over a year ago

thanks, it works. i have another question. Why the output when printed is [255 255 255] or [1 1 1], not [255, 255, 255] ?

yatu Over a year ago

Sorry I don't really understand what you mean. @IrchamAjiNugroho

ircham Over a year ago

sorry for my bad english and explanation :D i've found the answer

Gabio · Accepted Answer · 2020-04-07 08:25:06Z

1

try this one:

data = [
    [
        [255, 255, 255],
        [255, 255, 255],
        [0, 0, 0],
        [255, 255, 255]
    ],
    [
        [0, 0, 0],
        [255, 255, 255],
        [0, 0, 0],
        [0, 0, 0],
    ],
    [
        [255, 255, 255],
        [255, 255, 255],
        [0, 0, 0],
        [0, 0, 0],
    ]
]

result = [[1 if y == [255, 255, 255] else 0 for y in x] for x in data]
print(result)

Bare in mind that it will work if the only possible values are [255, 255, 255] or [0, 0, 0] (This is what I understand from the question)

edited Apr 7, 2020 at 8:25

answered Apr 7, 2020 at 8:10

Gabio

9,5643 gold badges17 silver badges38 bronze badges

5 Comments

Aven Desta Over a year ago

I would go with this answer.

yatu Over a year ago

Using lists and loopy approaches to work with numpy arrays is never the way to go if there are numpy functions to do the same. If you compare performances on larger arrays, you'll see huge differences

Gabio Over a year ago

It will be significant on large arrays as you said but the OP didn't mention the scale of his data so this is, IMHO, the simplest solution.

yatu Over a year ago

255 and 0s... This is for image data. Usual sizes are of the order of 1024x1024, which already would give you 1048576 different data points for a single image, and that is considering a single channel. For image data, using highly optimized numpy and scipy functions is crucial if you want to have a reasonably performant approach, which is by no means possible with python level loops.

ircham Over a year ago

thanks by the way, the size of an array of pixels of an image

DenverCoder1 · Accepted Answer · 2020-04-07 08:42:20Z

0

This is how I would do it:

Using tuple(col) == (255, 255, 255) is important because the truth value of an array with more than one element is ambiguous.

data = np.asarray(im) # (assuming this is how you got your data)

new_data = np.array(np.zeros((len(data),len(data[0]))))

for x,row in enumerate(data):
    for y,col in enumerate(row):
        if tuple(col) == (255, 255, 255):
            new_data[x][y] = 1

print(new_data)

Or in one line:

new_data = [[1 if tuple(col) == (255, 255, 255) else 0 for col in row] for row in data]

The first will result in a numpy array and the second, just a list. However, you can convert it to a numpy array by using:

new_data = np.array([[1 if tuple(col) == (255, 255, 255) else 0 for col in row] for row in data])

edited Apr 7, 2020 at 8:42

answered Apr 7, 2020 at 8:27

DenverCoder1

2,5211 gold badge12 silver badges27 bronze badges

Collectives™ on Stack Overflow

Change the content of the array in python

3 Answers 3

3 Comments

5 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

5 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related