Closed qin-yu closed 4 months ago
According to https://github.com/hci-unihd/plant-seg?tab=readme-ov-file#pre-trained-networks, step size 0.0001
should work for all existing models, i.e. to have 3 significant figures in resolutions of future models, I believe 4 decimal places should be enough, but let's use 0.00001
just in case.
E.g. if I have voxel size
[0.28, 0.06500000162500004, 0.06500000162500004]
or at least[0.28, 0.065, 0.065]
, I can only put 0.06 or 0.07 for Y and X.