Create an extended view of a numpy array

Question

I am learning image convolution and in order to handle the edges, I want to repeat the first and last rows/columns of the original array (at the center) like this :

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

(The values in the corners correspond to the diagonal elements.)

I will iterate over each pixel of the original array and use the array above to extract its 3x3 neighbors. So here is my question: is it possible to generate the view corresponding to this array so that I don't need to store a new (n+2)x(m+2) array in memory?

user2357112 · Accepted Answer · 2015-12-03 18:54:15Z

1

There's no way to do this as a view, since there's no way to make the strides work. For a not-view solution, numpy.pad does the job. The signal-processing routines you're using might also let you specify a padding type.

answered Dec 3, 2015 at 18:54

user2357112

286k32 gold badges490 silver badges569 bronze badges

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

1 Comment

John Over a year ago

Thank you for your quick answer! I was aware of scipy.signals.convolve2d and its "boundary" parameter but I need to code it myself. I will use numpy.pad and if memory becomes a problem, I will handle the edges manually in my loop.

Collectives™ on Stack Overflow

Create an extended view of a numpy array

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related