How to get the mean of a 3D array over 2D?

Question

I have an array with the following shape (7352, 128, 6)

How can I get the "mean" of the 1st and 2nd dimensions with NumPy in Python?

The resulting shape of the "means" I want to obtain is (1, 6).

`np.mean` has an `axis` parameter, which is exactly for this — Ander Biguri, Feb 25 '20 at 13:49

score 3 · Accepted Answer · answered Feb 25 '20 at 13:49

3

You can do:

np.mean(x, axis=(0, 1))

You resulting shape would be (6,) though.

answered Feb 25 '20 at 13:49

Thomas Schillaci

2,348
1
7
19

Thank you, this was the solution I was looking for and yeah, sorry you're right about the shape! – RabbitBadger Feb 25 '20 at 13:52

score 1 · Answer 2 · answered Feb 25 '20 at 13:52

1

numpy.mean() with axis argument is proper solution.

import numpy as np

x = np.random.rand(7352, 128, 6)
x_mean = np.mean(x, axis=(0, 1))

print(x_mean.shape)  # -> (6,)

answered Feb 25 '20 at 13:52

Filip Młynarski

3,534
1
10
22

score 1 · Answer 3 · answered Feb 25 '20 at 13:55

You can use the axis parameter of the mean function:

img = np.random.rand(7352,128,6)
y_mean = img.mean(axis=0) # mean over y (row)
x_mean = img.mean(axis=1) # mean over x (column)
z_mean = img.mean(axis=(0,1)) # the mean you want (shape is (6,))

This post may also help.

How to get the mean of a 3D array over 2D?

3 Answers3