Pythonic way of indexing 2d array with list of coordinates

Question

I have coordinates of class numpy.ndarray and of dimensions (200,2). Basically the output from regionprops.

I want to index a image matrix (also an ndarray) of dim img.shape = (1000,1000). I want to be able to do something like

for prop in region:
    img[prop.coords] = 0

However, what happens is, it doesn't see the coordinates as row, column pairs. But rather, each number as a row number and sets that entire row to 0.

How can I fix this?

I tried reshaping the array but that doesn't seem to work either. The only other think I can think of is to convert those indices into a matrix, where the matrix is same size as image dimensions and those coordinates are 1 and all the other coordinate values are 0. Then, index using this matrix. However, this doesn't seem any more efficient than just brute forcing it using a for loop.

Are you looking for img[prop.coords[:,0], prop.coords[:,1]] = 0? — Ian
– Ian, Commented Apr 23, 2019 at 7:48

Tonechas · Accepted Answer · 2019-04-23 08:06:28Z

6

You have to supply the coordinates as a tuple in order to invoke basic indexing:

for prop in region:
    img[tuple(prop.coords.T)] = 0

answered Apr 23, 2019 at 8:06

Tonechas

13.8k16 gold badges52 silver badges85 bronze badges

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

Comments

yann · Accepted Answer · 2019-04-23 08:43:21Z

0

for prop in region:
    (min_row, min_col, max_row, max_col) = prop['bboxtuple']
    img[min_row:max_row, min_col:max_col] = 0

answered Apr 23, 2019 at 8:43

yann

6627 silver badges14 bronze badges

Collectives™ on Stack Overflow

Pythonic way of indexing 2d array with list of coordinates

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related