Was really hoping to register using my full resolution data, but there seems to be a scaling issue.
ITK version: 4.13.2
Greedy commit: abef6386366cd1a7d296cd8df0d76c9b3368aaab
I'm using an old commit. I tried a fresh pull but had installation issues and did not have time to trouble shoot them. Thought this issue might persist for newer versions and should document it.
The reslice works with a skip sampled version of the moving image mov_ss = mov[::2, ::2, ::4]:
I believe this issue also affects registrations. When using mov.nrrd as the moving image and 2 scale levels, the first scale level runs fine (image is smaller by a factor of 8 voxels), but seg faults when moving up to the full resolution iterations.
Hi Paul,
Was really hoping to register using my full resolution data, but there seems to be a scaling issue.
ITK version: 4.13.2 Greedy commit: abef6386366cd1a7d296cd8df0d76c9b3368aaab I'm using an old commit. I tried a fresh pull but had installation issues and did not have time to trouble shoot them. Thought this issue might persist for newer versions and should document it.
The reslice works with a skip sampled version of the moving image
mov_ss = mov[::2, ::2, ::4]:I believe this issue also affects registrations. When using
mov.nrrdas the moving image and 2 scale levels, the first scale level runs fine (image is smaller by a factor of 8 voxels), but seg faults when moving up to the full resolution iterations.