如何在 numpy 数组上应用掩码，如果掩码的值为 True，则保持原始值不变，如果为 False，则将其设置为零？

Question

I have a 3D numpy array:

image = np.random.random((2, 2, 3))

[[[0.01188816 0.46263957 0.00943777]
  [0.0742566  0.8375209  0.259363  ]]

 [[0.30823133 0.17924745 0.74292469]
  [0.68490255 0.03143513 0.68233715]]]

and a mask:

[[[ True  True  True]
  [False False False]]

 [[False False False]
  [False False False]]]

desired output:

[[[0.01188816 0.46263957 0.00943777]
  [0          0.         0.        ]]

 [[0.         0.         0.        ]
  [0.         0.         0.        ]]]

So, the original array should be modified according to the mask - if the mask's value is False, the entry should be set to 0, otherwise left unchanged.

What I tried:

(image[unknown_array])
[0.01188816 0.46263957 0.00943777]

This indeed gives the right values, but without the zeros. How can I get the zeros to the right place?

Thank you very much for any help.

Answer 1

You can do it using different ways:

x = np.arange(1,7).reshape(2,3)  # numbers from 1 to 6
mask = x % 2 == 0                # mask for even numbers
print(x, mask)
# (array([[1, 2, 3],
#         [4, 5, 6]]),
#  array([[False,  True, False],
#         [ True, False,  True]]))

Szczesny suggestion is perhaps the simplest one:

y = x * mask
print(y)
# array([[0, 2, 0],
#        [4, 0, 6]])

Filling the y array "by hand":

y = np.zeros_like(x)
y[mask] = x[mask]
print(y)
# array([[0, 2, 0],
#        [4, 0, 6]])