setting values on multiple indexes in a numpy array

Question

I have the following fragment of python code:

import numpy as np


# Some random array
x = np.random.rand(34, 176, 256)
# Get some indexes
pos_idx = np.argwhere(x > 0.5)
# Sample some values from these indexes
seeds = pos_idx[np.random.choice(pos_idx.shape[0], size=5, replace=False), :]
# Now create another array
y = np.zeros_like(x, np.uint8)
y[seeds] = 1

The last line given an error something along the lines of:

index 77 is out of bounds for axis 0 with size 34

But I am not sure how this can happen as all the sampled indexes should be valid as they are a subset.

The array produced by argwhere cannot be used directly to index an array. The output of np.where is designed for that. — hpaulj
– hpaulj, Commented Aug 1, 2020 at 21:45
You're right, @hpaulj. However, the OP has an extra twist on the problem: selecting a random sample of indexes that meet the condition, and only setting those elements in an array of the same shape as the input. How can one achieve that with the np.where approach? — Basil
– Basil, Commented Aug 2, 2020 at 1:13

Roohollah Etemadi · Accepted Answer · 2020-08-01 21:24:01Z

3

It considers the values in seeds as indexes of the first dimension. To access the elements by their indexes in seeds, you can use:

y[seeds[:,0],seeds[:,1],seeds[:,2]] = 1

answered Aug 1, 2020 at 21:24

Roohollah Etemadi

1,4531 gold badge10 silver badges19 bronze badges

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

Comments

Alpha Green · Accepted Answer · 2020-08-01 21:37:19Z

1

This code will help you setting the values to 1

import numpy as np

# Some random array
x = np.random.rand(34, 176, 256)
# Get some indexes
pos_idx = np.argwhere(x > 0.5)
# Sample some values from these indexes
seeds = pos_idx[np.random.choice(pos_idx.shape[0], size=5, replace=False), :]
# Now create another array
y = np.zeros_like(x, np.uint8)
for i in seeds:
    y[tuple(i)] = 1

answered Aug 1, 2020 at 21:37

Alpha Green

10811 bronze badges

Comments

jsmart · Accepted Answer · 2020-08-01 20:53:56Z

0

Does seeds have the right shape? If I understand correctly, x has random values; y has the same shape as x but is all zeros; and seeds is some index values, and these locations will be set to one.

print('x    : ', x.ndim, x.shape)
print('y    : ', y.ndim, y.shape)
print('seeds: ', seeds.ndim, seeds.shape)

x    :  3 (34, 176, 256)
y    :  3 (34, 176, 256)
seeds:  2 (5, 3)

answered Aug 1, 2020 at 20:53

jsmart

3,0111 gold badge9 silver badges14 bronze badges

1 Comment

Luca Over a year ago

Each element in seeds is a 3 dimensional coordinates/index and I am basically trying to set that index to 1 in y

Collectives™ on Stack Overflow

setting values on multiple indexes in a numpy array

3 Answers 3

Comments

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related