def _run_interface(self, runtime):
## Load the 4D image files
img = nb.load(self.inputs.in_file)
data = img.get_data()
affine = img.get_affine()
## Load the gradient strengths and directions
bvals = np.loadtxt(self.inputs.bvals)
gradients = np.loadtxt(self.inputs.bvecs).T
## Place in Dipy's preferred format
gtab = GradientTable(gradients)
gtab.bvals = bvals
## Mask the data so that tensors are not fit for
## unnecessary voxels
mask = data[..., 0] > 50
## Fit the tensors to the data
tenmodel = dti.TensorModel(gtab)
tenfit = tenmodel.fit(data, mask)
## Calculate the mode of each voxel's tensor
mode_data = tenfit.mode
## Write as a 3D Nifti image with the original affine
img = nb.Nifti1Image(mode_data, affine)
out_file = op.abspath(self._gen_outfilename())
nb.save(img, out_file)
iflogger.info('Tensor mode image saved as {i}'.format(i=out_file))
return runtime
def _run_interface(self, runtime):
## Load the 4D image files
img = nb.load(self.inputs.in_file)
data = img.get_data()
affine = img.get_affine()
## Load the gradient strengths and directions
bvals = np.loadtxt(self.inputs.bvals)
gradients = np.loadtxt(self.inputs.bvecs).T
## Place in Dipy's preferred format
gtab = GradientTable(gradients)
gtab.bvals = bvals
## Mask the data so that tensors are not fit for
## unnecessary voxels
mask = data[..., 0] > 50
## Fit the tensors to the data
tenmodel = dti.TensorModel(gtab)
tenfit = tenmodel.fit(data, mask)
## Calculate the mode of each voxel's tensor
mode_data = tenfit.mode
## Write as a 3D Nifti image with the original affine
img = nb.Nifti1Image(mode_data, affine)
out_file = op.abspath(self._gen_outfilename())
nb.save(img, out_file)
iflogger.info('Tensor mode image saved as {i}'.format(i=out_file))
return runtime
def nonlinfit_fn(dwi, bvecs, bvals, base_name):
import nibabel as nb
import numpy as np
import os.path as op
import dipy.reconst.dti as dti
from dipy.core.gradients import GradientTable
dwi_img = nb.load(dwi)
dwi_data = dwi_img.get_data()
dwi_affine = dwi_img.get_affine()
from dipy.segment.mask import median_otsu
b0_mask, mask = median_otsu(dwi_data, 2, 4)
# Mask the data so that tensors are not fit for
# unnecessary voxels
mask_img = nb.Nifti1Image(mask.astype(np.float32), dwi_affine)
b0_imgs = nb.Nifti1Image(b0_mask.astype(np.float32), dwi_affine)
b0_img = nb.four_to_three(b0_imgs)[0]
out_mask_name = op.abspath(base_name + '_binary_mask.nii.gz')
out_b0_name = op.abspath(base_name + '_b0_mask.nii.gz')
nb.save(mask_img, out_mask_name)
nb.save(b0_img, out_b0_name)
# Load the gradient strengths and directions
bvals = np.loadtxt(bvals)
gradients = np.loadtxt(bvecs)
# Dipy wants Nx3 arrays
if gradients.shape[0] == 3:
gradients = gradients.T
assert(gradients.shape[1] == 3)
# Place in Dipy's preferred format
gtab = GradientTable(gradients)
gtab.bvals = bvals
# Fit the tensors to the data
tenmodel = dti.TensorModel(gtab, fit_method="NLLS")
tenfit = tenmodel.fit(dwi_data, mask)
# Calculate the fit, fa, and md of each voxel's tensor
tensor_data = tenfit.lower_triangular()
print('Computing anisotropy measures (FA, MD, RGB)')
from dipy.reconst.dti import fractional_anisotropy, color_fa
evals = tenfit.evals.astype(np.float32)
FA = fractional_anisotropy(np.abs(evals))
FA = np.clip(FA, 0, 1)
MD = dti.mean_diffusivity(np.abs(evals))
norm = dti.norm(tenfit.quadratic_form)
RGB = color_fa(FA, tenfit.evecs)
evecs = tenfit.evecs.astype(np.float32)
mode = tenfit.mode.astype(np.float32)
mode = np.nan_to_num(mode)
# Write tensor as a 4D Nifti image with the original affine
tensor_fit_img = nb.Nifti1Image(tensor_data.astype(np.float32), dwi_affine)
mode_img = nb.Nifti1Image(mode.astype(np.float32), dwi_affine)
norm_img = nb.Nifti1Image(norm.astype(np.float32), dwi_affine)
FA_img = nb.Nifti1Image(FA.astype(np.float32), dwi_affine)
evecs_img = nb.Nifti1Image(evecs, dwi_affine)
evals_img = nb.Nifti1Image(evals, dwi_affine)
rgb_img = nb.Nifti1Image(np.array(255 * RGB, 'uint8'), dwi_affine)
MD_img = nb.Nifti1Image(MD.astype(np.float32), dwi_affine)
out_tensor_file = op.abspath(base_name + "_tensor.nii.gz")
out_mode_file = op.abspath(base_name + "_mode.nii.gz")
out_fa_file = op.abspath(base_name + "_fa.nii.gz")
out_norm_file = op.abspath(base_name + "_norm.nii.gz")
out_evals_file = op.abspath(base_name + "_evals.nii.gz")
out_evecs_file = op.abspath(base_name + "_evecs.nii.gz")
out_rgb_fa_file = op.abspath(base_name + "_rgb_fa.nii.gz")
out_md_file = op.abspath(base_name + "_md.nii.gz")
nb.save(rgb_img, out_rgb_fa_file)
nb.save(norm_img, out_norm_file)
nb.save(mode_img, out_mode_file)
nb.save(tensor_fit_img, out_tensor_file)
nb.save(evecs_img, out_evecs_file)
nb.save(evals_img, out_evals_file)
nb.save(FA_img, out_fa_file)
nb.save(MD_img, out_md_file)
print('Tensor fit image saved as {i}'.format(i=out_tensor_file))
print('FA image saved as {i}'.format(i=out_fa_file))
print('MD image saved as {i}'.format(i=out_md_file))
return out_tensor_file, out_fa_file, out_md_file, \
out_evecs_file, out_evals_file, out_rgb_fa_file, out_norm_file, \
out_mode_file, out_mask_name, out_b0_name
def nonlinfit_fn(dwi, bvecs, bvals, base_name):
import nibabel as nb
import numpy as np
import os.path as op
import dipy.reconst.dti as dti
from dipy.core.gradients import GradientTable
dwi_img = nb.load(dwi)
dwi_data = dwi_img.get_data()
dwi_affine = dwi_img.get_affine()
from dipy.segment.mask import median_otsu
b0_mask, mask = median_otsu(dwi_data, 2, 4)
# Mask the data so that tensors are not fit for
# unnecessary voxels
mask_img = nb.Nifti1Image(mask.astype(np.float32), dwi_affine)
b0_imgs = nb.Nifti1Image(b0_mask.astype(np.float32), dwi_affine)
b0_img = nb.four_to_three(b0_imgs)[0]
out_mask_name = op.abspath(base_name + '_binary_mask.nii.gz')
out_b0_name = op.abspath(base_name + '_b0_mask.nii.gz')
nb.save(mask_img, out_mask_name)
nb.save(b0_img, out_b0_name)
# Load the gradient strengths and directions
bvals = np.loadtxt(bvals)
gradients = np.loadtxt(bvecs).T
# Place in Dipy's preferred format
gtab = GradientTable(gradients)
gtab.bvals = bvals
# Fit the tensors to the data
tenmodel = dti.TensorModel(gtab, fit_method="NLLS")
tenfit = tenmodel.fit(dwi_data, mask)
# Calculate the fit, fa, and md of each voxel's tensor
tensor_data = tenfit.lower_triangular()
print('Computing anisotropy measures (FA, MD, RGB)')
from dipy.reconst.dti import fractional_anisotropy, color_fa
evals = tenfit.evals.astype(np.float32)
FA = fractional_anisotropy(np.abs(evals))
FA = np.clip(FA, 0, 1)
MD = dti.mean_diffusivity(np.abs(evals))
norm = dti.norm(tenfit.quadratic_form)
RGB = color_fa(FA, tenfit.evecs)
evecs = tenfit.evecs.astype(np.float32)
mode = tenfit.mode.astype(np.float32)
# Write tensor as a 4D Nifti image with the original affine
tensor_fit_img = nb.Nifti1Image(tensor_data.astype(np.float32), dwi_affine)
mode_img = nb.Nifti1Image(mode.astype(np.float32), dwi_affine)
norm_img = nb.Nifti1Image(norm.astype(np.float32), dwi_affine)
FA_img = nb.Nifti1Image(FA.astype(np.float32), dwi_affine)
evecs_img = nb.Nifti1Image(evecs, dwi_affine)
evals_img = nb.Nifti1Image(evals, dwi_affine)
rgb_img = nb.Nifti1Image(np.array(255 * RGB, 'uint8'), dwi_affine)
MD_img = nb.Nifti1Image(MD.astype(np.float32), dwi_affine)
out_tensor_file = op.abspath(base_name + "_tensor.nii.gz")
out_mode_file = op.abspath(base_name + "_mode.nii.gz")
out_fa_file = op.abspath(base_name + "_fa.nii.gz")
out_norm_file = op.abspath(base_name + "_norm.nii.gz")
out_evals_file = op.abspath(base_name + "_evals.nii.gz")
out_evecs_file = op.abspath(base_name + "_evecs.nii.gz")
out_rgb_fa_file = op.abspath(base_name + "_rgb_fa.nii.gz")
out_md_file = op.abspath(base_name + "_md.nii.gz")
nb.save(rgb_img, out_rgb_fa_file)
nb.save(norm_img, out_norm_file)
nb.save(mode_img, out_mode_file)
nb.save(tensor_fit_img, out_tensor_file)
nb.save(evecs_img, out_evecs_file)
nb.save(evals_img, out_evals_file)
nb.save(FA_img, out_fa_file)
nb.save(MD_img, out_md_file)
print('Tensor fit image saved as {i}'.format(i=out_tensor_file))
print('FA image saved as {i}'.format(i=out_fa_file))
print('MD image saved as {i}'.format(i=out_md_file))
return out_tensor_file, out_fa_file, out_md_file, \
out_evecs_file, out_evals_file, out_rgb_fa_file, out_norm_file, \
out_mode_file, out_mask_name, out_b0_name
