Calculate mean of each image block - Python

Question

I have a matrix like below

c = [[ 1 2 3 4 5 6 7 8 9 1]
     [ 2 3 4 5 6 7 8 9 1 2]
     [ 3 4 5 6 7 8 9 1 2 3]
     [ 4 5 6 7 8 9 1 2 3 4]]

From the given SO ANSWERS in this post, I used it to divide the matrix into blocks (2*5) like below

def blockshaped(arr, nrows, ncols):
"""
Return an array of shape (n, nrows, ncols) where
n * nrows * ncols = arr.size

If arr is a 2D array, the returned array should look like n subblocks with
each subblock preserving the "physical" layout of arr.
"""
    h, w = arr.shape
    assert h % nrows == 0, "{} rows is not evenly divisble by {}".format(h, nrows)
    assert w % ncols == 0, "{} cols is not evenly divisble by {}".format(w, ncols)
    return (arr.reshape(h//nrows, nrows, -1, ncols)
           .swapaxes(1,2)
           .reshape(-1, nrows, ncols))


print(blockshaped(c, 2, 5))

Result:

[[[ 1 2 3 4 5 ]
  [ 2 3 4 5 6 ]]

 [[ 6 7 8 9 1 ]
  [ 7 8 9 1 2]]

 [[ 3 4 5 6 7 ]
  [ 4 5 6 7 8 ]]

 [[ 8 9 1 2 3 ]
  [ 9 1 2 3 4]]]

I got 4 blocks of matrix, and now I need the mean value of each block. How to calculate the mean of this each block?

When I try to use the mean() it is going to calculate the mean for the whole matrix but not for each block.

For the functionality provided by `blockshaped()`, check out [`skimage.util.view_as_blocks()`](https://scikit-image.org/docs/dev/api/skimage.util.html#view-as-blocks). — AMC, Dec 13 '19 at 10:22

seralouk · Accepted Answer · 2019-12-13T09:50:21.147

1. Half line solution with List comprehension

results = blockshaped(c, 2, 5)
block_means = [np.mean(results[block,:,:]) for block in range(results.shape[0])]

print(block_means)
# [3.5, 5.8, 5.5, 4.2]

Version 2 -- shorter code:

results = blockshaped(c, 2, 5)
block_means = [np.mean(block) for block in results]
# [3.5, 5.8, 5.5, 4.2]

In [15]: %timeit [np.mean(results[block,:,:]) for block in range(results.shape[0])]
10000 loops, best of 3: 35.9 µs per loop

In [16]: %timeit [np.mean(block) for block in results]
10000 loops, best of 3: 33.4 µs per loop

P.S: The second solution will only work if the blocks are in the first (0) dimension of results.

Can you benchmark these on larger inputs? Measuring such tiny runtimes isn’t particularly accurate or useful IMO. — AMC, Dec 13 '19 at 10:23

score 1 · Answer 2 · edited Dec 13 '19 at 09:55

1

Another option is to use the map function:

means = np.round(map(np.mean, r),3)
print(means)

which results in:

[ 3.5 5.8 5.5 4.2]

edited Dec 13 '19 at 09:55

seralouk

30,938
9
118
133

answered Dec 13 '19 at 09:43

David

8,113
2
17
36

Calculate mean of each image block - Python

2 Answers2

1. Half line solution with List comprehension

Version 2 -- shorter code: