CorpusCallosum.registration.midsagittal_plane_alignment

class CorpusCallosum.registration.midsagittal_plane_alignment.MidplaneDebugVolumes(distance_diff, candidate_mask, left_mask, right_mask)

Debug volumes produced by distance-map midplane refinement.

Attributes

distance_diff	(np.ndarray) Voxel-wise difference d_left - d_right in fsaverage space.
candidate_mask	(np.ndarray) Boolean mask selecting voxels used for plane fitting.
left_mask	(np.ndarray) Partitioned binary left-hemisphere mask in fsaverage space.
right_mask	(np.ndarray) Partitioned binary right-hemisphere mask in fsaverage space.

class CorpusCallosum.registration.midsagittal_plane_alignment.MidplaneRefinementResult(updated_vox2vox, center_shift_vox, diagnostics, debug_volumes)

Container for distance-map refinement outputs.

Attributes

updated_vox2vox	(AffineMatrix4x4) Updated subject-to-fsaverage voxel transform.
center_shift_vox	(float) Estimated LR center shift in fsaverage voxels.
diagnostics	(dict[str, object]) JSON-serializable diagnostics and rejection reasons.
debug_volumes	(MidplaneDebugVolumes) Optional intermediate volumes for quality control and debugging.

class CorpusCallosum.registration.midsagittal_plane_alignment.MidplaneTransformResult(orig2fsavg_vox2vox, fsavg_vox2ras, fsavg_header_dict, fsavg_shape, base_middle_vox, midline_shift_vox, midline_shift_diagnostics)

Result returned by find_midplane_transform().

Attributes

orig2fsavg_vox2vox	(AffineMatrix4x4) Subject-to-fsaverage voxel transform after optional refinement.
fsavg_vox2ras	(AffineMatrix4x4) fsaverage voxel-to-RAS transform used for downstream mapping.
fsavg_header_dict	(MGHHeaderDict) Header metadata for the fsaverage-aligned volume.
fsavg_shape	(Shape3d) Shape of the fsaverage-aligned target grid.
base_middle_vox	(float) Baseline midsagittal x-coordinate in fsaverage voxel units.
midline_shift_vox	(float) Applied LR shift (or estimated center shift) in voxel units.
midline_shift_diagnostics	(dict[str, object]) Method-specific diagnostics for refinement decisions.

CorpusCallosum.registration.midsagittal_plane_alignment.find_midplane_transform(orig, aseg_img, midplane_method)

Resolve the fsaverage midplane transform for a selected refinement method.

Parameters:

orignibabelImage

Input conformed anatomical image.

aseg_imgnibabelImage

Subject aseg segmentation image.

midplane_methodstr

Midplane strategy. Supported values are "center", "fsaverage", "fsaverage_distance_map", and "fsaverage_symmetry".

Returns:

MidplaneTransformResult

Transform bundle and method-specific diagnostics for downstream CC steps.

CorpusCallosum.registration.midsagittal_plane_alignment.refine_midline_lr_shift(orig2fsavg_vox2vox, aseg_data, fsavg_shape, base_middle_vox, max_shift_vox=6, step_vox=1)

Refine midsagittal alignment by searching small LR translations.

Parameters:

orig2fsavg_vox2voxAffineMatrix4x4

Initial subject-to-fsaverage voxel transform.

aseg_datanp.ndarray

Subject segmentation labels in native voxel space.

fsavg_shapeShape3d

Shape of the fsaverage target grid.

base_middle_voxfloat

Baseline midsagittal x-coordinate in fsaverage voxels.

max_shift_voxint, default=6

Maximum absolute LR shift explored.

step_voxint, default=1

Integer step size for the LR search.

Returns:

tuple[AffineMatrix4x4, int, float, dict[str, object]]

Updated transform, selected shift, selected score, and diagnostics.

Notes

Candidate shifts are regularized and validated with conservative rejection criteria (boundary hits, low support, and weak improvements over baseline).

CorpusCallosum.registration.midsagittal_plane_alignment.refine_midplane_with_distance_maps(orig2fsavg_vox2vox, aseg_data, fsavg_shape, base_middle_vox, zero_band_vox=0.5, support_distance_vox=24.0, fit_band_vox=20.0, max_tilt_deg=7.5, max_center_shift_vox=8.0)

Fit a midsagittal plane from LR distance-map symmetry in fsaverage space.

Parameters:

orig2fsavg_vox2voxAffineMatrix4x4

Initial subject-to-fsaverage voxel transform.

aseg_datanp.ndarray

Subject segmentation labels in native voxel space.

fsavg_shapeShape3d

Shape of the fsaverage target grid.

base_middle_voxfloat

Baseline midsagittal x-coordinate in fsaverage voxels.

zero_band_voxfloat, default=0.5

Absolute distance-difference threshold for candidate midplane voxels.

support_distance_voxfloat, default=24.0

Maximum distance to either hemisphere mask for candidate support.

fit_band_voxfloat, default=20.0

X-range around baseline center used for plane fitting.

max_tilt_degfloat, default=7.5

Maximum accepted angular tilt away from the LR axis.

max_center_shift_voxfloat, default=8.0

Maximum accepted LR center shift from baseline.

Returns:

MidplaneRefinementResult

Updated transform, applied center shift, diagnostics, and debug volumes.

CorpusCallosum.registration.midsagittal_plane_alignment.register_centroids_to_fsavg(aseg_nib)

Estimate a rigid subject-to-fsaverage alignment from aseg centroids.

Parameters:

aseg_nibnibabelImage

Input aseg image in subject space.

Returns:

tuple[AffineMatrix4x4, AffineMatrix4x4, AffineMatrix4x4, MGHHeaderDict]

(aseg2fsavg_vox2vox, aseg2fsavg_ras2ras, fsavg_hires_vox2ras, fsavg_header).

CorpusCallosum.registration.midsagittal_plane_alignment.resample_segmentation_to_fsavg(seg_data, orig2fsavg_vox2vox, fsavg_shape)

Resample a segmentation into fsaverage voxel space.

Parameters:

seg_datanp.ndarray

Input segmentation array in subject voxel space.

orig2fsavg_vox2voxAffineMatrix4x4

Subject-to-fsaverage voxel transform.

fsavg_shapeShape3d

Output shape in fsaverage voxel coordinates.

Returns:

np.ndarray

Resampled integer segmentation in fsaverage space.

Notes

Nearest-neighbor interpolation is used to preserve discrete labels.