Efficient summed Area Table Calculation with Numpy

Question

I\'m trying to calculate a summed area table of a feature count matrix using python and numpy. Currently I\'m using the following code:

def summed_area_table         


        

Answer 1

There's a NumPy function cumsum for cumulative sums. Applying it twice yields the desired table:

import numpy as np

A = np.random.randint(0, 10, (3, 4))

print A
print A.cumsum(axis=0).cumsum(axis=1)

Output:

[[7 4 7 2]
 [6 9 9 5]
 [6 6 7 6]]
[[ 7 11 18 20]
 [13 26 42 49]
 [19 38 61 74]]

Performance analysis: (https://stackoverflow.com/a/25351344/3419103)

import numpy as np
import time

A = np.random.randint(0, 10, (3200, 1400))

t = time.time()
S = A.cumsum(axis=0).cumsum(axis=1)
print np.round_(time.time() - t, 3), 'sec elapsed'

Output:

0.15 sec elapsed