Add nlsam denoising modules

[ThoumyreStanislas]

For nlsam, the script requires a bet to be made before, to give a brain mask, so the question is: do we make a module for nlsam and put the bet inside the module, or do we take the bet from the betcrop module? And secondly, is it really worth adding nlsam to nf-neuro? However, the nslam denoising is very efficient, but it takes a little time and above all, we don't want to give the non bet image, because otherwise it takes too much time.