we need to be careful of cells on the borders of patches, because their nearest tumor areas might be in adjacent patches. use "halo" or "ghost" cells. we can probably implement this by taking a patch and padding it by the tumor microenvironment distance. we don't consider cells in the padded region, but we do consider tumor boundaries. we need to be careful when setting the stride... let's use this issue to figure out that stride value.
we need to be careful of cells on the borders of patches, because their nearest tumor areas might be in adjacent patches. use "halo" or "ghost" cells. we can probably implement this by taking a patch and padding it by the tumor microenvironment distance. we don't consider cells in the padded region, but we do consider tumor boundaries. we need to be careful when setting the stride... let's use this issue to figure out that stride value.