mlh3789 mlh3789 - 2 months ago 20
Python Question

Quick way to upsample numpy array by nearest neighbor tiling

I have a 2D array of integers that is MxN, and I would like to expand the array to (BM)x(BN) where B is the length of a square tile side thus each element of the input array is repeated as a BxB block in the final array. Below is an example with a nested for loop. Is there a quicker/builtin way?

import numpy as np

a = np.arange(9).reshape([3,3]) # input array - 3x3
B=2. # block size - 2
A = np.zeros([a.shape[0]*B,a.shape[1]*B]) # output array - 6x6

# Loop, filling A with tiled values of a at each index
for i,l in enumerate(a): # lines in a
for j,aij in enumerate(l): # a[i,j]
A[B*i:B*(i+1),B*j:B*(j+1)] = aij


Result ...

a= [[0 1 2]
[3 4 5]
[6 7 8]]

A = [[ 0. 0. 1. 1. 2. 2.]
[ 0. 0. 1. 1. 2. 2.]
[ 3. 3. 4. 4. 5. 5.]
[ 3. 3. 4. 4. 5. 5.]
[ 6. 6. 7. 7. 8. 8.]
[ 6. 6. 7. 7. 8. 8.]]

YXD YXD
Answer Source

One option is

>>> a.repeat(2, axis=0).repeat(2, axis=1)
array([[0, 0, 1, 1, 2, 2],
       [0, 0, 1, 1, 2, 2],
       [3, 3, 4, 4, 5, 5],
       [3, 3, 4, 4, 5, 5],
       [6, 6, 7, 7, 8, 8],
       [6, 6, 7, 7, 8, 8]])

This is slightly wasteful due to the intermediate array but it's concise at least.