How to extract only the pixels of an image where it is masked? (Python numpy array operation)

Question

I have an image and its corresponding mask for the cob as numpy arrays:

The image numpy array has shape (332, 107, 3).

The mask is Boolean (consists of True/False) and has this shape as binary (332, 107).

 [[False False False ... False False False]
 [False False False ... False False False]
 [False False False ... False False False]
 ...
 [False False False ... False False False]
 [False False False ... False False False]
 [False False False ... False False False]]

How can I get the color pixels of the cob (all pixels in the color image where the mask is)?

What have you tried? A simple conditional should do it, test for the colour of the mask if the test passes, add the pixels from the image. — Felipe Gutierrez
– Felipe Gutierrez, Commented Dec 3, 2019 at 15:28
I am new at programming. I tried setting all pixels of the image which are not in the mask at 0: img[mask==False]=0 but this did not work. You mean I should loop over every element in the array? — Olympia
– Olympia, Commented Dec 3, 2019 at 15:32
if I use img[mask[..., None]] I get this error: IndexError: boolean index did not match indexed array along dimension 2; dimension is 3 but corresponding boolean dimension is 1 — Olympia
– Olympia, Commented Dec 3, 2019 at 15:34
I think you want cob = img * mask because False evaluates to zero. — Mark Setchell
– Mark Setchell, Commented Dec 3, 2019 at 15:42

Olympia · Accepted Answer · 2019-12-04 08:12:58Z

6

Thanks to the useful comment of M.Setchell, I was able to find the answer myself.

Basically, I had to expand the dimensions of the mask array (2D) to the same dimension of the image (3D with 3 color channels).

y=np.expand_dims(mask,axis=2)
newmask=np.concatenate((y,y,y),axis=2)

Then I had to simply multiply the new mask with the image to get the colored mask:

cob= img * newmask

And here just for visualization the result:

answered Dec 4, 2019 at 8:12

Olympia

5571 gold badge8 silver badges19 bronze badges

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

1 Comment

Michael Litvin Over a year ago

This doesn't actually answer the question of how to extract the pixels...

yazan sayed · Accepted Answer · 2022-11-26 21:43:29Z

0

If you want to get an array of the pixels, i.e. array with shape (n,3):

#assuming mask.shape = (h,w) , and mask.dtype = bool
pixels = img[[mask]]

and if you want to produce the image in your answer then simply do this:

cop = img.copy()
cop[mask] = 0

edited Nov 26, 2022 at 21:43

answered Nov 26, 2022 at 21:18

yazan sayed

1,1391 gold badge14 silver badges27 bronze badges

Collectives™ on Stack Overflow

How to extract only the pixels of an image where it is masked? (Python numpy array operation)

2 Answers 2

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related