Here is my Problem:
I have a bunch of images where I want to extend the canvas size to either 1024x1024, 2048x2048 or 4096x4096.
Here is how I like it to go:
- read image
- look for largest width or height
- automaticly decide which to use, 1024x1024, 2048x2048 or 4096x4096 (the best fit one)
- expand_y / expand_x
If there is another method without expand_ that would be fine to, the important part for me is that expand has the “neumann boundary conditions” which just fills the rest of the canvas with “smeared” pixels.