Python loops too slow for numpy ndarray

Question

I have an image as a matrix (ndarray) of shape (720, 1280, 3) with RGB pixels (variable: original) and another ndarray of shape (720, 1280) which is made of booleans (variable: im).

for every True value in im I want original's corresponding pixel to be of color [0, 0, 255].

I tried (with both arrays flattened)

for i in range(im.size):
  if(im[i] == True):
    original[i] = [0, 0, 255]

but its too slow to have it as a video output.

Any ideas how I can speed this up?

From array's perspective, code (1D) and comments (2D) don't match. — CristiFati
– CristiFati, Commented Apr 30, 2021 at 19:19

Camilo Martínez M. · Accepted Answer · 2021-04-30 19:25:22Z

5

This is the same as doing (without flattening your arrays):

original[im == True] = [0, 0, 255]

That is the vectorized representation of your for loop.

answered Apr 30, 2021 at 19:25

Camilo Martínez M.

1,6291 gold badge10 silver badges23 bronze badges

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

1 Comment

GoodDeeds Over a year ago

Or just original[im] = [0,0,255].

Collectives™ on Stack Overflow

Python loops too slow for numpy ndarray

1 Answer 1

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Related