Skip to content

Commit

Permalink
Merge pull request #671 from karanphil/add_variance_to_vizualize_fodf
Browse files Browse the repository at this point in the history 
Adding variance to scil_visualize_fodf
  • Loading branch information
@arnaudbore
arnaudbore authored Mar 7, 2023
2 parents 711e389 + 7f50d71 commit 749f9f6
Show file tree
Hide file tree
Showing 2 changed files with 104 additions and 26 deletions.
55 changes: 47 additions & 8 deletions scilpy/viz/scene_utils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@
import numpy as np

import vtk
from dipy.reconst.shm import sh_to_sf_matrix
from dipy.reconst.shm import sh_to_sf_matrix, sh_to_sf
from fury import window, actor
from fury.colormap import distinguishable_colormap
from PIL import Image
Expand Down Expand Up @@ -122,7 +122,8 @@ def set_display_extent(slicer_actor, orientation, volume_shape, slice_index):

def create_odf_slicer(sh_fodf, orientation, slice_index, mask, sphere,
nb_subdivide, sh_order, sh_basis, full_basis,
scale, radial_scale, norm, colormap):
scale, radial_scale, norm, colormap, sh_variance=None,
variance_k=1, variance_color=(255, 255, 255)):
"""
Create a ODF slicer actor displaying a fODF slice. The input volume is a
3-dimensional grid containing the SH coefficients of the fODF for each
Expand Down Expand Up @@ -158,6 +159,12 @@ def create_odf_slicer(sh_fodf, orientation, slice_index, mask, sphere,
If True, enables normalization of ODF slicer.
colormap : str
Colormap for the ODF slicer. If None, a RGB colormap is used.
sh_variance : np.ndarray, optional
Spherical harmonics of the variance fODF data.
variance_k : float, optional
Factor that multiplies sqrt(variance).
variance_color : tuple, optional
Color of the variance fODF data, in RGB.
Returns
-------
Expand All @@ -172,14 +179,46 @@ def create_odf_slicer(sh_fodf, orientation, slice_index, mask, sphere,
B_mat = sh_to_sf_matrix(sphere, sh_order, sh_basis,
full_basis, return_inv=False)

odf_actor = actor.odf_slicer(sh_fodf, mask=mask, norm=norm,
radial_scale=radial_scale,
sphere=sphere,
colormap=colormap,
scale=scale, B_matrix=B_mat)
var_actor = None

if sh_variance is not None:
fodf = sh_to_sf(sh_fodf, sphere, sh_order, sh_basis,
full_basis=full_basis)
fodf_var = sh_to_sf(sh_variance, sphere, sh_order, sh_basis,
full_basis=full_basis)
fodf_uncertainty = fodf + variance_k * np.sqrt(np.clip(fodf_var, 0,
None))
# normalise fodf and variance
if norm:
maximums = np.abs(np.append(fodf, fodf_uncertainty, axis=-1))\
.max(axis=-1)
fodf[maximums > 0] /= maximums[maximums > 0][..., None]
fodf_uncertainty[maximums > 0] /= maximums[maximums > 0][..., None]

odf_actor = actor.odf_slicer(fodf, mask=mask, norm=False,
radial_scale=radial_scale,
sphere=sphere, scale=scale,
colormap=colormap)

var_actor = actor.odf_slicer(fodf_uncertainty, mask=mask, norm=False,
radial_scale=radial_scale,
sphere=sphere, scale=scale,
colormap=variance_color)
var_actor.GetProperty().SetDiffuse(0.0)
var_actor.GetProperty().SetAmbient(1.0)
var_actor.GetProperty().SetFrontfaceCulling(True)
else:
odf_actor = actor.odf_slicer(sh_fodf, mask=mask, norm=norm,
radial_scale=radial_scale,
sphere=sphere,
colormap=colormap,
scale=scale, B_matrix=B_mat)
set_display_extent(odf_actor, orientation, sh_fodf.shape[:3], slice_index)
if var_actor is not None:
set_display_extent(var_actor, orientation,
fodf_uncertainty.shape[:3], slice_index)

return odf_actor
return odf_actor, var_actor


def _get_affine_for_texture(orientation, offset):
Expand Down
75 changes: 57 additions & 18 deletions scripts/scil_visualize_fodf.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -100,52 +100,72 @@ def _build_arg_parser():
help='Disable normalization of ODF slicer.')

# Background image options
p.add_argument('--background',
bg = p.add_argument_group('Background arguments')
bg.add_argument('--background',
help='Background image file. If RGB, values must '
'be between 0 and 255.')

p.add_argument('--bg_range', nargs=2, metavar=('MIN', 'MAX'), type=float,
bg.add_argument('--bg_range', nargs=2, metavar=('MIN', 'MAX'), type=float,
help='The range of values mapped to range [0, 1] '
'for background image. [(bg.min(), bg.max())]')

p.add_argument('--bg_opacity', type=float, default=1.0,
bg.add_argument('--bg_opacity', type=float, default=1.0,
help='The opacity of the background image. Opacity of 0.0 '
'means transparent and 1.0 is completely visible. '
'[%(default)s]')

p.add_argument('--bg_offset', type=float, default=0.5,
bg.add_argument('--bg_offset', type=float, default=0.5,
help='The offset of the background image. [%(default)s]')

p.add_argument('--bg_interpolation',
bg.add_argument('--bg_interpolation',
default='nearest', choices={'linear', 'nearest'},
help='Interpolation mode for the background image. '
'[%(default)s]')

p.add_argument('--bg_color', nargs=3, type=float, default=(0, 0, 0),
bg.add_argument('--bg_color', nargs=3, type=float, default=(0, 0, 0),
help='The color of the overall background, behind '
'everything. Must be RGB values scaled between 0 and '
'1. [%(default)s]')

# Peaks input file options
p.add_argument('--peaks',
peaks = p.add_argument_group('Peaks arguments')
peaks.add_argument('--peaks',
help='Peaks image file.')

p.add_argument('--peaks_color', nargs=3, type=float,
peaks.add_argument('--peaks_color', nargs=3, type=float,
help='Color used for peaks, as RGB values scaled between 0 '
'and 1. If None, then a RGB colormap is used. '
'[%(default)s]')

p.add_argument('--peaks_width', default=1.0, type=float,
peaks.add_argument('--peaks_width', default=1.0, type=float,
help='Width of peaks segments. [%(default)s]')

peaks_scale_group = p.add_mutually_exclusive_group()
peaks_scale = p.add_argument_group('Peaks scaling arguments', 'Choose '
'between peaks values and arbitrary '
'length.')
peaks_scale_group = peaks_scale.add_mutually_exclusive_group()
peaks_scale_group.add_argument('--peaks_values',
help='Peaks values file.')

peaks_scale_group.add_argument('--peaks_length', default=0.65, type=float,
help='Length of the peaks segments. '
'[%(default)s]')

# fODF variance options
var = p.add_argument_group('Variance arguments', 'For the visualization '
'of fodf uncertainty, the variance is used '
'as follow: mean + k * sqrt(variance), where '
'mean is the input fodf (in_fodf) and k is the '
'scaling factor (variance_k).')
var.add_argument('--variance',
help='FODF variance file.')
var.add_argument('--variance_k', default=1, type=float,
help='Scaling factor (k) for the computation of the fodf '
'uncertainty. [%(default)s]')
var.add_argument('--var_color', nargs=3, type=float, default=(1, 1, 1),
help='Color of variance outline. Must be RGB values scaled '
'between 0 and 1. [%(default)s]')

return p


Expand Down Expand Up @@ -222,6 +242,16 @@ def _get_data_from_inputs(args):
peak_vals =\
nib.nifti1.load(args.peaks_values).get_fdata(dtype=np.float32)
data['peaks_values'] = peak_vals
if args.variance:
assert_same_resolution([args.variance, args.in_fodf])
variance = nib.nifti1.load(args.variance).get_fdata(dtype=np.float32)
if len(variance.shape) == 3:
variance = np.reshape(variance, variance.shape + (1,))
if variance.shape != fodf.shape:
raise ValueError('Dimensions mismatch between fODF {0} and '
'variance {1}.'
.format(fodf.shape, variance.shape))
data['variance'] = variance

return data

Expand All @@ -246,17 +276,26 @@ def main():
else:
color_rgb = None

variance = data['variance'] if args.variance else None
var_color = np.asarray(args.var_color) * 255
# Instantiate the ODF slicer actor
odf_actor = create_odf_slicer(data['fodf'], args.axis_name,
args.slice_index, mask, sph,
args.sph_subdivide, sh_order,
args.sh_basis, full_basis,
args.scale,
not args.radial_scale_off,
not args.norm_off,
args.colormap or color_rgb)
odf_actor, var_actor = create_odf_slicer(data['fodf'], args.axis_name,
args.slice_index, mask, sph,
args.sph_subdivide, sh_order,
args.sh_basis, full_basis,
args.scale,
not args.radial_scale_off,
not args.norm_off,
args.colormap or color_rgb,
sh_variance=variance,
variance_k=args.variance_k,
variance_color=var_color)
actors.append(odf_actor)

# Instantiate a variance slicer actor if a variance image is supplied
if 'variance' in data:
actors.append(var_actor)

# Instantiate a texture slicer actor if a background image is supplied
if 'bg' in data:
bg_actor = create_texture_slicer(data['bg'],
Expand Down

0 comments on commit 749f9f6

Please sign in to comment.