Element extraction from multiple arrays in Numpy/Scipy

Question

I have three numpy arrays: A, B, and C. I have to extract '1' that is common (pixel wise) to any two of the given arrays, and set all other elements as 0.

import numpy as np

A = np.array([[1,0,0,0,0],
               [1,0,0,0,0],
               [1,0,0,0,0]])              

B = np.array([[1,1,1,1,1],
               [1,1,1,1,1],
               [1,1,1,1,1]])

C = np.array([[0,0,0,0,1],
               [0,0,0,0,1],
               [0,0,0,0,1]])

The expected answer would be:

result = np.array([[1,0,0,0,1],
               [1,0,0,0,1],
               [1,0,0,0,1]])

How can I do it using Numpy/Scipy? A faster performance is necessary since in my real problem, the number of arrays are 50s and the size of each array is (3000, 3000), and I have to extract the '1' if common to 30 arrays.

result = (A & B) | (A & C) | (B & C)

Wör Du Schnaffzig
– Wör Du Schnaffzig

2020-10-12 12:09:03 +00:00
Commented Oct 12, 2020 at 12:09 — Wör Du Schnaffzig
– Wör Du Schnaffzig, Commented Oct 12, 2020 at 12:09

Ffisegydd · Accepted Answer · 2015-01-19 11:35:52Z

2

Sum them and then use boolean operations to check whether the elements are >=2.

import numpy as np

A = np.array([[1,0,0,0,0],
               [1,0,0,0,0],
               [1,0,0,0,0]])              

B = np.array([[1,1,1,1,1],
               [1,1,1,1,1],
               [1,1,1,1,1]])

C = np.array([[0,0,0,0,1],
               [0,0,0,0,1],
               [0,0,0,0,1]])

D = A + B + C

D = (D >= 2).astype(int)

print(D)
#[[1 0 0 0 1]
# [1 0 0 0 1]
# [1 0 0 0 1]]

(D >= 2).astype(int) will return True for each value that is >=2 and False otherwise. You then convert it from True/False values to 1/0 values using .astype(int).

You mention in your question that in your real case you could have up to 50 arrays? Either store them in a list and use sum() or store them in a 3-dimensional numpy array and use your_numpy_array.sum().

answered Jan 19, 2015 at 11:35

Ffisegydd

54k16 gold badges151 silver badges125 bronze badges

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

2 Comments

Borys Over a year ago

Thanks for your answer. Does it really work for my 50 arrays when extracting common in any 30 arrays as follows: (D >= 30).astype(int)?

Borys Over a year ago

The shape of my array in the real case is (50L, 3000L, 3000L). How can I apply axis?

Collectives™ on Stack Overflow

Element extraction from multiple arrays in Numpy/Scipy

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related