Make dilate and erode accept floats

[ErichZimmer]

Hi all,

I was using this library as a lightweight alternative to scipy, but I ran into a few inconveniences. I usually work with 2D images that are 32 bit floats and normalized to [0..1]. However, it seems some functions, notably erode and dilate, do not accept floating point arrays. Is there a particular reason to why erode and dilate only accepts integer data types? If so, it is no biggie as I can simply emulate floating point arithmetics with 32 bit integers.

Kind regards,
Erich

[luispedro]

I don't know if there is a particular reason, besides this originally being written to replicate other integer-only code that was much slower. Looking at the code, it may even just work well as is by removing the current limitations

[ErichZimmer]

When getting accustomed to this library's code base, I ran into something interesting on line 117. To my assumption, this is the cause of unexpected results on the function erode. Is there a particular reason why erode_sub is written the way it is?

[luispedro]

The original reason was probably to make it behave like pymorph: at the time, pymorph did a lot of what I wanted, but was Python-only and 2D only, so a lot of mahotas was written to be a C++ version of that

[ErichZimmer]

Should I implement dilate and erode functions that use a boolean footprint like what is implemented in SciPy? Currently, the phase separation algorithm I am implementing does not work with the current dilate and erode functions but works for SciPy and my own implementation.

[luispedro]

Yes, I'd be happy to review/help.