include gputools.deconv as deconvolution backend about lattice_lightsheet_deskew_deconv

[VolkerH]

a quick hack to try this out https://github.com/VolkerH/Lattice_Lightsheet_Deskew_Deconv/blob/gputools_deconv/lls_dd/deconvolution_gputools.py
Worked surprisingly well.
Signficantly slower than flowdec, but also significantly faster than CPU and not limited to CUDA.
Does not seem to keep the FFT plan, so significant potential for speedup by keeping this.
Doesn't throw any memory allocation errors for datasets where flowdec competes with gputools.
Also has a spatial deconvolutoin option (without FFT). Might be worth a try for small PSFs if we can also find a seperable approximation.

Todo:

• create an abstract deconvolver class with an init and a process method.
• have concrete child classes for flowdec and gputools. Can also add cudaDeconv backend, but that requires different processing of PSFs so is slightly more complicated. Conversion of the PSF could be integrated into the init method of the deconvolver rather than modifying the current PSF handling.

[VolkerH]

closing this as this is implemented now. Not yet as a generic class that wraps all deconvolution backends but working.

[VolkerH]

Works reasonably fast with the changes in: https://github.com/VolkerH/Lattice_Lightsheet_Deskew_Deconv/blob/master/lls_dd/deconv_gputools_rewrite.py

Visually not quite the same quality as the flowdec result, but probably not too hard to include the additional tweaks. Low priority at the moment.