Python - Average values of numpy multidimensionnal array every specific time

Question

I have a multidimensionnal array "test[:,:,:]" and i would like to get averaged values on the test.shape[0] dimension for every 4 "frames" i would like to keep the same dimensions of my array and substitute the 4 values by the mean value.

As example:

test=np.array([[[ 2.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 3.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 3.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 5.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 2.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 3.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 3.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 5.,  1.,  1.],
        [ 1.,  1.,  1.]],

       [[ 2.,  1.,  1.],
        [ 1.,  1.,  1.]]])        


for i in range(test.shape[0]-1,4):
    test_mean = (test[i,:,:]+test[i+1,:,:]+test[i+2,:,:]+test[i+3,:,:])/4.

But, i don't keep the same dimension...what is the best way to do that?

It is not very clear what you mean by "I don't keep the same dimension". So, for the (9, 2, 3) array test, do you want a (9, 2, 3) array as a result, a (2, 2, 3) array, or something else? We could help you best if you could explicitly add the correct result for test_mean for reference. — Stefan
– Stefan, Commented Jul 21, 2014 at 11:24
It is also a good idea to stay online for a while after posting a question. Otherwise potential answerers might move on if they can't get you to comment on requests for clarification. — Stefan
– Stefan, Commented Jul 21, 2014 at 11:28
As i would say, i would like to keep the (9, 2, 3) array! Thanks ;) (& i had to go). — user3601754
– user3601754, Commented Jul 21, 2014 at 11:31

Stefan · Accepted Answer · 2014-07-21 11:13:33Z

2

You are overwriting test_mean every time. A good start is:

test_mean = np.zeros_like(test)
for i in xrange(test.shape[0]-4):
    test_mean[i] = test[i:i+4].mean(axis=0)

Here is a more efficient implementation from scipy:

from scipy.ndimage import uniform_filter1d
test_mean2 = uniform_filter1d(test, 4, axis=0)

Check the documentation to understand how the result is stored and what options you have to treat boundary values.

edited Jul 21, 2014 at 11:13

answered Jul 21, 2014 at 11:03

Stefan

4,6102 gold badges33 silver badges34 bronze badges

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

Collectives™ on Stack Overflow

Python - Average values of numpy multidimensionnal array every specific time

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related