CODEWITHSUNDEEP
pythonnumpy
Talha Yousuf
Talha Yousuf

Reputation: 301

How to split multiple masks from a numpy array in single line?

I have a 500,500,3 shaped numpy array containing repetitive integer values. I can use np.unique to get unique integer values present in the said array (because they can be different for every such array).

Is there any way to split multiple masks in one line.

Here masks is a numpy array. It contains different repetitive integer values

 ids = np.unique(masks) # ids = [0, 1, 2] for example
 # currently doing this
 mask0 = masks == ids[0]
 mask1 = masks == ids[1]
 mask2 = masks == ids[2]

Is there any single line method to get a set of all binary masks. For example something like.

 all_masks = masks == ids[:] # for example

Upvotes: 0

Views: 861

Answers (1)

Ollin Boer Bohan
Ollin Boer Bohan

Reputation: 2401

You can broadcast both arrays to a shape of (id, height, width, channels), and then check equality:

all_masks = (masks[np.newaxis] == ids[:, np.newaxis, np.newaxis, np.newaxis])

In the result, all_masks[i] is the (height, width, channels) binary mask for the ith id.

Upvotes: 1

Related Questions