split numpy multidimensional array into equal pieces

Asked Feb 20 '17 at 15:51

Active Feb 20 '17 at 16:23

Viewed 66 times

I'm trying to split a 9x512 image (called test) into 384 3x4 images, but trying the following code:

splited = test.reshape(-1,3,4)
plt.imshow(splited[0,:,:])

I don't get the same result as doing:

plt.imshow(test[0:3,0:4])

I would really appreciate your help :)

edited Feb 20 '17 at 16:23

demongolem

asked Feb 20 '17 at 15:51

florpi

okay, but you need to tell us which approach gives the "correct" result? – MSeifert Feb 20 '17 at 16:04
check this question http://stackoverflow.com/questions/16856788/slice-2d-array-into-smaller-2d-arrays – Yohanes Gultom Feb 20 '17 at 16:04
the correct must be plt.imshow(test[0:3,0:4]), but I'm trying to avoid for loops since I have too many images – florpi Feb 20 '17 at 16:06
well, you do `imshow` on your images, that will introduce much more overhead than any python-iterations. – MSeifert Feb 20 '17 at 16:12
@YohanesGultom that was what I was looking for, thank you! :) – florpi Feb 20 '17 at 16:23
So it looks like you want `test..reshape(3,3,128,4)`, and possibly a transpose to produce a (3,128,3,4) and reshape to (384,3,4)? – hpaulj Feb 20 '17 at 17:24

0 Answers0