foa3d.odf

foa3d.odf.compute_odf_map(fbr_vec, px_sz, odf, odi, fbr_dnst, vec_tnsr_eig, scale, norm, deg=6, vx_thr=0.5, vec_thr=1e-06)

Compute the spherical harmonics coefficients iterating over super-voxels of fiber orientation vectors.

Parameters

• fbr_vec (NumPy memory-map object (axis order=(Z,Y,X,C), dtype=float)) – fiber orientation vectors

• px_sz (numpy.ndarray (shape=(3,), dtype=float)) – adjusted isotropic pixel size [μm]

• odf (NumPy memory-map object (axis order=(X,Y,Z,C), dtype=float32)) – initialized array of ODF spherical harmonics coefficients

• odi (dict) –

  orientation dispersion dictionary

  odi_pri: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  primary orientation dispersion index

  odi_sec: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  secondary orientation dispersion index

  odi_tot: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  total orientation dispersion index

  odi_anis: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  orientation dispersion index anisotropy

• fbr_dnst (NumPy memory-map object (axis order=(Z,Y,X), dtype=float)) – initialized fiber density image

• vec_tnsr_eig (NumPy memory-map object (axis order=(Z,Y,X,C), dtype=float32)) – initialized array of orientation tensor eigenvalues

• scale (int) – side of the ODF super-voxel [px]

• norm (numpy.ndarray (dtype: float)) – 2D array of spherical harmonics normalization factors

• deg (int) – degrees of the spherical harmonics series expansion

• vx_thr (float) – minimum relative threshold on the sliced voxel volume

• vec_thr (float) – minimum relative threshold on non-zero orientation vectors

Returns

• odf (numpy.ndarray (axis order=(X,Y,Z,C), dtype=float32)) – volumetric map of real-valued spherical harmonics coefficients

• fbr_dnst (numpy.ndarray (axis order=(Z,Y,X), dtype=float)) – fiber density image

foa3d.odf.compute_orientation_dispersion(vec_tnsr_eig, odi_pri, odi_sec, odi_tot, odi_anis)

Compute orientation dispersion parameters.

Parameters

• vec_tnsr_eig (NumPy memory-map object (axis order=(Z,Y,X,C), dtype=float32)) – orientation tensor eigenvalues computed from an ODF super-voxel

• odi_pri (NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)) – primary orientation dispersion index

• odi_sec (NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)) – secondary orientation dispersion index

• odi_tot (NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)) – total orientation dispersion index

• odi_anis (NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)) – orientation dispersion index anisotropy

Return type

None

foa3d.odf.compute_vec_tensor_eigen(fbr_vec)

Compute the eigenvalues of the 3x3 orientation tensor obtained from a reshaped super-voxel of fiber orientation vectors.

Parameters

fbr_vec (numpy.ndarray (shape=(N,3), dtype=float)) – fiber orientation vectors (reshaped super-voxel of shape=(Nz,Ny,Nx), i.e. N=Nz*Ny*Nx)

Returns

vec_tensor_eigen – orientation tensor eigenvalues in ascending order

Return type

numpy.ndarray (shape=(3,), dtype=float32)

foa3d.odf.generate_odf_background(bg_mrtrix, fbr_vec, scale, iso_fbr=None)

Generate the down-sampled background image required to visualize the 3D ODF map in MRtrix3.

Parameters

• bg_mrtrix (numpy.ndarray (axis order=(X,Y,Z), dtype=uint8)) – initialized background array for ODF visualization in MRtrix3

• fbr_vec (numpy.ndarray (axis order=(Z,Y,X,C), dtype=float32)) – fiber orientation vectors

• scale (int) – side of the ODF super-voxel [px]

• iso_fbr (numpy.ndarray (axis order=(Z,Y,X), dtype=uint8)) – isotropic fiber image

Return type

None

foa3d.odf.init_odf_arrays(vec_img_shp, tmp_dir, scale, deg=6, exp_all=False)

Initialize the output datasets of the ODF analysis stage.

Parameters

• vec_img_shp (tuple) – vector volume shape [px]

• tmp_dir (str) – path to temporary folder

• scale (int) – fiber ODF resolution (super-voxel side [px])

• deg (int) – degrees of the spherical harmonics series expansion

• exp_all (bool) – export all images

Returns

• odf (NumPy memory-map object (axis order=(X,Y,Z,C), dtype=float32)) – initialized array of ODF spherical harmonics coefficients

• odi (dict) –

  dispersion image dictionary

  odi_pri: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  initialized array of primary orientation dispersion parameters

  odi_sec: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  initialized array of secondary orientation dispersion parameters

  odi_tot: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  initialized array of total orientation dispersion parameters

  odi_anis: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  initialized array of orientation dispersion anisotropy parameters

• fbr_dnst (NumPy memory-map object (axis order=(Z,Y,X), dtype=float)) – initialized fiber density image

• bg_mrtrix (NumPy memory-map object (axis order=(X,Y,Z), dtype=uint8)) – initialized background for ODF visualization in MRtrix3

• vec_tensor_eigen (NumPy memory-map object (axis order=(Z,Y,X,C), dtype=float32)) – initialized fiber orientation tensor eigenvalues

foa3d.odf.mask_orientation_dispersion(vec_tnsr_eig, odi)

Suppress orientation dispersion background.

Parameters

• vec_tnsr_eig (NumPy memory-map object (axis order=(Z,Y,X,C), dtype=float32)) – orientation tensor eigenvalues computed from an ODF super-voxel

• odi (dict) –

  orientation dispersion dictionary

  odi_pri: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  array of primary orientation dispersion parameters

  odi_sec: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  array of secondary orientation dispersion parameters

  odi_tot: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  array of total orientation dispersion parameters

  odi_anis: NumPy memory-map object (axis order=(Z,Y,X), dtype=float32)

  array of orientation dispersion anisotropy parameters

Return type

None