Skip to content

Fixes to the anisotropic segmentation of FastSurfer#695

Merged
m-reuter merged 3 commits into
Deep-MI:devfrom
dkuegler:fix/aniso
May 19, 2025
Merged

Fixes to the anisotropic segmentation of FastSurfer#695
m-reuter merged 3 commits into
Deep-MI:devfrom
dkuegler:fix/aniso

Conversation

@dkuegler
Copy link
Copy Markdown
Member

Several fixes to voxelsize-associated settings in FastSurfer

  • conform did not support int datatype voxel sizes, which only happen when manually passed.
  • Fix CerebNet bounding boxes that are too small (smaller than the patch size).
  • Fix when CerebNet was complaining about voxel sizes (inverted check)
  • Fix brun_fastsurfer.sh issue: brun did not properly assign the device in some cases (when multiple devices were passed).
  • Fix viewagg_device issue, where auto always defaulted to device number 0 instead of to the device set with --device (for VINN, CerebNet, HypVINN).
  • Extract function orientation_to_ornts from to_target_orientation for consistency, also introduce StrictOrientationType
  • Add better output to is_conform that also shows the current situation (image size and voxel size and affine) to the output
  • Fix the passing of the voxel size (zoom) to FastSurferVINN and HypVINN
  • Some formatting fixes

…hen manually passed.

Fix CerebNet bounding boxes that are too small (smaller than the patch size).
Fix when CerebNet was complaining about voxel sizes (inverted check)
@dkuegler dkuegler requested a review from m-reuter May 19, 2025 16:14
… in some cases (when multiple devices were passed).

Fix viewagg_device issue, where auto always defaulted to device number 0 instead of to the device set with --device (for VINN, CerebNet, HypVINN).
Extract function orientation_to_ornts from to_target_orientation for consistency, also introduce StrictOrientationType
Add better output to is_conform that also shows the current situation (image size and voxel size and affine) to the output
Fix the passing of the voxel size (zoom) to FastSurferVINN and HypVINN
Some formatting fixes
Remove anisotropy TODOs
@m-reuter m-reuter merged commit 0a694c5 into Deep-MI:dev May 19, 2025
3 checks passed
@dkuegler dkuegler deleted the fix/aniso branch May 20, 2025 09:03
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants