I am using dwi2mask for some neonatal dwi data and is it doing a pretty good job, but I am interested in seeing if there are any suggestions for improving it (so I have to do less manual correction). Or is there another tool (perhaps dhcp?) that would work? Thank you!
I usually use bet for masking the neonatal data, and I don’t perform manual edits. For the major part of the processes, I think that the masks doesn’t need to be perfectly precise. When I need more accurate masks, what I do is (as you suggested) propagate the mask obtained from the T2w using the dHCP protocol.