segment

Module: segment.benchmarks.bench_quickbundles

Benchmarks for QuickBundles

Run all benchmarks with:

import dipy.segment as dipysegment
dipysegment.bench()


With Pytest, Run this benchmark with:

pytest -svv -c bench.ini /path/to/bench_quickbundles.py

 MDFpy Attributes Metric Computes a distance between two sequential data. QB_New Streamlines alias of nibabel.streamlines.array_sequence.ArraySequence assert_array_equal(x, y[, err_msg, verbose]) Raises an AssertionError if two array_like objects are not equal. assert_arrays_equal(arrays1, arrays2) assert_equal(actual, desired[, err_msg, verbose]) Raises an AssertionError if two objects are not equal. get_fnames([name]) Provide full paths to example or test datasets. load_tractogram(filename, reference[, …]) Load the stateful tractogram from any format (trk, tck, vtk, fib, dpy) measure(code_str[, times, label]) Return elapsed time for executing code in the namespace of the caller. set_number_of_points Change the number of points of streamlines

Module: segment.bundles

 BundleMinDistanceAsymmetricMetric([num_threads]) Asymmetric Bundle-based Minimum distance BundleMinDistanceMetric([num_threads]) Bundle-based Minimum Distance aka BMD BundleSumDistanceMatrixMetric([num_threads]) Bundle-based Sum Distance aka BMD RecoBundles(streamlines[, greater_than, …]) Methods StreamlineLinearRegistration([metric, x0, …]) Methods Streamlines alias of nibabel.streamlines.array_sequence.ArraySequence chain chain(*iterables) –> chain object apply_affine(aff, pts) Apply affine matrix aff to points pts ba_analysis(recognized_bundle, expert_bundle) Calculates bundle adjacency score between two given bundles bundle_adjacency(dtracks0, dtracks1, threshold) Find bundle adjacency between two given tracks/bundles bundle_shape_similarity(bundle1, bundle2, rng) Calculates bundle shape similarity between two given bundles using bundle adjacency (BA) metric bundles_distances_mam Calculate distances between list of tracks A and list of tracks B bundles_distances_mdf Calculate distances between list of tracks A and list of tracks B check_range(streamline, gt, lt) cluster_bundle(bundle, clust_thr, rng[, …]) Clusters bundles length Euclidean length of streamlines nbytes(streamlines) qbx_and_merge(streamlines, thresholds[, …]) Run QuickBundlesX and then run again on the centroids of the last layer Select a random set of streamlines set_number_of_points Change the number of points of streamlines Return the current time in seconds since the Epoch.

Module: segment.clustering

 ABCMeta Metaclass for defining Abstract Base Classes (ABCs). AveragePointwiseEuclideanMetric Computes the average of pointwise Euclidean distances between two sequential data. Cluster([id, indices, refdata]) Provides functionalities for interacting with a cluster. ClusterCentroid(centroid[, id, indices, refdata]) Provides functionalities for interacting with a cluster. ClusterMap([refdata]) Provides functionalities for interacting with clustering outputs. ClusterMapCentroid([refdata]) Provides functionalities for interacting with clustering outputs that have centroids. Clustering Methods Identity Provides identity indexing functionality. Metric Computes a distance between two sequential data. MinimumAverageDirectFlipMetric Computes the MDF distance (minimum average direct-flip) between two sequential data. QuickBundles(threshold[, metric, …]) Clusters streamlines using QuickBundles [Garyfallidis12]. QuickBundlesX(thresholds[, metric]) Clusters streamlines using QuickBundlesX. ResampleFeature Extracts features from a sequential datum. TreeCluster(threshold, centroid[, indices]) Attributes TreeClusterMap(root) Attributes abstractmethod(funcobj) A decorator indicating abstract methods. nbytes(streamlines) qbx_and_merge(streamlines, thresholds[, …]) Run QuickBundlesX and then run again on the centroids of the last layer set_number_of_points Change the number of points of streamlines Return the current time in seconds since the Epoch.

Module: segment.mask

 applymask(vol, mask) Mask vol with mask. binary_dilation(input[, structure, …]) Multidimensional binary dilation with the given structuring element. Compute the bounding box of nonzero intensity voxels in the volume. clean_cc_mask(mask) Cleans a segmentation of the corpus callosum so no random pixels are included. color_fa(fa, evecs) Color fractional anisotropy of diffusion tensor crop(vol, mins, maxs) Crops the input volume. fractional_anisotropy(evals[, axis]) Return Fractional anisotropy (FA) of a diffusion tensor. generate_binary_structure(rank, connectivity) Generate a binary structure for binary morphological operations. median_filter(input[, size, footprint, …]) Calculate a multidimensional median filter. median_otsu(input_volume[, vol_idx, …]) Simple brain extraction tool method for images from DWI data. multi_median(input, median_radius, numpass) Applies median filter multiple times on input data. otsu(image[, nbins]) Return threshold value based on Otsu’s method. segment_from_cfa(tensor_fit, roi, threshold) Segment the cfa inside roi using the values from threshold as bounds. warn(/, message[, category, stacklevel, source]) Issue a warning, or maybe ignore it or raise an exception.

Module: segment.metric

 ArcLengthFeature Extracts features from a sequential datum. AveragePointwiseEuclideanMetric Computes the average of pointwise Euclidean distances between two sequential data. CenterOfMassFeature Extracts features from a sequential datum. CosineMetric Computes the cosine distance between two vectors. EuclideanMetric alias of dipy.segment.metricspeed.SumPointwiseEuclideanMetric Feature Extracts features from a sequential datum. IdentityFeature Extracts features from a sequential datum. Metric Computes a distance between two sequential data. MidpointFeature Extracts features from a sequential datum. MinimumAverageDirectFlipMetric Computes the MDF distance (minimum average direct-flip) between two sequential data. ResampleFeature Extracts features from a sequential datum. SumPointwiseEuclideanMetric Computes the sum of pointwise Euclidean distances between two sequential data. VectorOfEndpointsFeature Extracts features from a sequential datum. dist Computes a distance between datum1 and datum2. distance_matrix Computes the distance matrix between two lists of sequential data. mdf(s1, s2) Computes the MDF (Minimum average Direct-Flip) distance [Garyfallidis12] between two streamlines.

Module: segment.threshold

 otsu(image[, nbins]) Return threshold value based on Otsu’s method. upper_bound_by_percent(data[, percent]) Find the upper bound for visualization of medical images upper_bound_by_rate(data[, rate]) Adjusts upper intensity boundary using rates

Module: segment.tissue

 ConstantObservationModel Observation model assuming that the intensity of each class is constant. IteratedConditionalModes Methods TissueClassifierHMRF([save_history, verbose]) This class contains the methods for tissue classification using the Markov Random Fields modeling approach add_noise(signal, snr, S0[, noise_type]) Add noise of specified distribution to the signal from a single voxel.

MDFpy

class dipy.segment.benchmarks.bench_quickbundles.MDFpy

Bases: dipy.segment.metricspeed.Metric

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible(shape1, shape2) Checks if features can be used by metric.dist based on their shape. dist(features1, features2) Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

are_compatible(shape1, shape2)

Checks if features can be used by metric.dist based on their shape.

Basically this method exists so we don’t have to do this check inside the metric.dist function (speedup).

Parameters
shape1int, 1-tuple or 2-tuple

shape of the first data point’s features

shape2int, 1-tuple or 2-tuple

shape of the second data point’s features

Returns
are_compatiblebool

whether or not shapes are compatible

dist(features1, features2)

Computes a distance between two data points based on their features.

Parameters
features12D array

Features of the first data point.

features22D array

Features of the second data point.

Returns
double

Distance between two data points.

Metric

class dipy.segment.benchmarks.bench_quickbundles.Metric

Bases: object

Computes a distance between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between extracted features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

When subclassing Metric, one only needs to override the dist and are_compatible methods.

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

are_compatible()

Checks if features can be used by metric.dist based on their shape.

Basically this method exists so we don’t have to do this check inside the metric.dist function (speedup).

Parameters
shape1int, 1-tuple or 2-tuple

shape of the first data point’s features

shape2int, 1-tuple or 2-tuple

shape of the second data point’s features

Returns
are_compatiblebool

whether or not shapes are compatible

dist()

Computes a distance between two data points based on their features.

Parameters
features12D array

Features of the first data point.

features22D array

Features of the second data point.

Returns
double

Distance between two data points.

feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

QB_New

dipy.segment.benchmarks.bench_quickbundles.QB_New

Streamlines

dipy.segment.benchmarks.bench_quickbundles.Streamlines

alias of nibabel.streamlines.array_sequence.ArraySequence

assert_array_equal

dipy.segment.benchmarks.bench_quickbundles.assert_array_equal(x, y, err_msg='', verbose=True)

Raises an AssertionError if two array_like objects are not equal.

Given two array_like objects, check that the shape is equal and all elements of these objects are equal (but see the Notes for the special handling of a scalar). An exception is raised at shape mismatch or conflicting values. In contrast to the standard usage in numpy, NaNs are compared like numbers, no assertion is raised if both objects have NaNs in the same positions.

The usual caution for verifying equality with floating point numbers is advised.

Parameters
xarray_like

The actual object to check.

yarray_like

The desired, expected object.

err_msgstr, optional

The error message to be printed in case of failure.

verbosebool, optional

If True, the conflicting values are appended to the error message.

Raises
AssertionError

If actual and desired objects are not equal.

assert_allclose

Compare two array_like objects for equality with desired relative and/or absolute precision.

assert_array_almost_equal_nulp, assert_array_max_ulp, assert_equal

Notes

When one of x and y is a scalar and the other is array_like, the function checks that each element of the array_like object is equal to the scalar.

Examples

The first assert does not raise an exception:

>>> np.testing.assert_array_equal([1.0,2.33333,np.nan],
...                               [np.exp(0),2.33333, np.nan])


Assert fails with numerical imprecision with floats:

>>> np.testing.assert_array_equal([1.0,np.pi,np.nan],
...                               [1, np.sqrt(np.pi)**2, np.nan])
Traceback (most recent call last):
...
AssertionError:
Arrays are not equal

Mismatched elements: 1 / 3 (33.3%)
Max absolute difference: 4.4408921e-16
Max relative difference: 1.41357986e-16
x: array([1.      , 3.141593,      nan])
y: array([1.      , 3.141593,      nan])


Use assert_allclose or one of the nulp (number of floating point values) functions for these cases instead:

>>> np.testing.assert_allclose([1.0,np.pi,np.nan],
...                            [1, np.sqrt(np.pi)**2, np.nan],
...                            rtol=1e-10, atol=0)


As mentioned in the Notes section, assert_array_equal has special handling for scalars. Here the test checks that each value in x is 3:

>>> x = np.full((2, 5), fill_value=3)
>>> np.testing.assert_array_equal(x, 3)


assert_arrays_equal

dipy.segment.benchmarks.bench_quickbundles.assert_arrays_equal(arrays1, arrays2)

assert_equal

dipy.segment.benchmarks.bench_quickbundles.assert_equal(actual, desired, err_msg='', verbose=True)

Raises an AssertionError if two objects are not equal.

Given two objects (scalars, lists, tuples, dictionaries or numpy arrays), check that all elements of these objects are equal. An exception is raised at the first conflicting values.

When one of actual and desired is a scalar and the other is array_like, the function checks that each element of the array_like object is equal to the scalar.

This function handles NaN comparisons as if NaN was a “normal” number. That is, AssertionError is not raised if both objects have NaNs in the same positions. This is in contrast to the IEEE standard on NaNs, which says that NaN compared to anything must return False.

Parameters
actualarray_like

The object to check.

desiredarray_like

The expected object.

err_msgstr, optional

The error message to be printed in case of failure.

verbosebool, optional

If True, the conflicting values are appended to the error message.

Raises
AssertionError

If actual and desired are not equal.

Examples

>>> np.testing.assert_equal([4,5], [4,6])
Traceback (most recent call last):
...
AssertionError:
Items are not equal:
item=1
ACTUAL: 5
DESIRED: 6


The following comparison does not raise an exception. There are NaNs in the inputs, but they are in the same positions.

>>> np.testing.assert_equal(np.array([1.0, 2.0, np.nan]), [1, 2, np.nan])


bench_quickbundles

dipy.segment.benchmarks.bench_quickbundles.bench_quickbundles()

get_fnames

dipy.segment.benchmarks.bench_quickbundles.get_fnames(name='small_64D')

Provide full paths to example or test datasets.

Parameters
namestr

the filename/s of which dataset to return, one of: - ‘small_64D’ small region of interest nifti,bvecs,bvals 64 directions - ‘small_101D’ small region of interest nifti, bvecs, bvals

101 directions

• ‘aniso_vox’ volume with anisotropic voxel size as Nifti

• ‘fornix’ 300 tracks in Trackvis format (from Pittsburgh Brain Competition)

• ‘gqi_vectors’ the scanner wave vectors needed for a GQI acquisitions of 101 directions tested on Siemens 3T Trio

• ‘small_25’ small ROI (10x8x2) DTI data (b value 2000, 25 directions)

• ‘test_piesno’ slice of N=8, K=14 diffusion data

• ‘reg_c’ small 2D image used for validating registration

• ‘reg_o’ small 2D image used for validation registration

• ‘cb_2’ two vectorized cingulum bundles

Returns
fnamestuple

filenames for dataset

Examples

>>> import numpy as np
>>> from dipy.data import get_fnames
>>> fimg, fbvals, fbvecs = get_fnames('small_101D')
>>> data.shape == (6, 10, 10, 102)
True
>>> bvals.shape == (102,)
True
>>> bvecs.shape == (102, 3)
True


dipy.segment.benchmarks.bench_quickbundles.load_tractogram(filename, reference, to_space=<Space.RASMM: 'rasmm'>, to_origin=<Origin.NIFTI: 'center'>, bbox_valid_check=True, trk_header_check=True)

Load the stateful tractogram from any format (trk, tck, vtk, fib, dpy)

Parameters
filenamestring

Filename with valid extension

referenceNifti or Trk filename, Nifti1Image or TrkFile, Nifti1Header or

trk.header (dict), or ‘same’ if the input is a trk file. Reference that provides the spatial attribute. Typically a nifti-related object from the native diffusion used for streamlines generation

to_spaceEnum (dipy.io.stateful_tractogram.Space)

to_originEnum (dipy.io.stateful_tractogram.Origin)

NIFTI standard, default (center of the voxel) TRACKVIS standard (corner of the voxel)

bbox_valid_checkbool

Verification for negative voxel coordinates or values above the volume dimensions. Default is True, to enforce valid file.

Verification that the reference has the same header as the spatial attributes as the input tractogram when a Trk is loaded

Returns
outputStatefulTractogram

The tractogram to load (must have been saved properly)

measure

dipy.segment.benchmarks.bench_quickbundles.measure(code_str, times=1, label=None)

Return elapsed time for executing code in the namespace of the caller.

The supplied code string is compiled with the Python builtin compile. The precision of the timing is 10 milli-seconds. If the code will execute fast on this timescale, it can be executed many times to get reasonable timing accuracy.

Parameters
code_strstr

The code to be timed.

timesint, optional

The number of times the code is executed. Default is 1. The code is only compiled once.

labelstr, optional

A label to identify code_str with. This is passed into compile as the second argument (for run-time error messages).

Returns
elapsedfloat

Total elapsed time in seconds for executing code_str times times.

Examples

>>> times = 10
>>> etime = np.testing.measure('for i in range(1000): np.sqrt(i**2)', times=times)
>>> print("Time for a single execution : ", etime / times, "s")
Time for a single execution :  0.005 s


set_number_of_points

dipy.segment.benchmarks.bench_quickbundles.set_number_of_points()
Change the number of points of streamlines

(either by downsampling or upsampling)

Change the number of points of streamlines in order to obtain nb_points-1 segments of equal length. Points of streamlines will be modified along the curve.

Parameters
streamlinesndarray or a list or dipy.tracking.Streamlines

If ndarray, must have shape (N,3) where N is the number of points of the streamline. If list, each item must be ndarray shape (Ni,3) where Ni is the number of points of streamline i. If dipy.tracking.Streamlines, its common_shape must be 3.

nb_pointsint

integer representing number of points wanted along the curve.

Returns
new_streamlinesndarray or a list or dipy.tracking.Streamlines

Results of the downsampling or upsampling process.

Examples

>>> from dipy.tracking.streamline import set_number_of_points
>>> import numpy as np


One streamline, a semi-circle:

>>> theta = np.pi*np.linspace(0, 1, 100)
>>> x = np.cos(theta)
>>> y = np.sin(theta)
>>> z = 0 * x
>>> streamline = np.vstack((x, y, z)).T
>>> modified_streamline = set_number_of_points(streamline, 3)
>>> len(modified_streamline)
3


Multiple streamlines:

>>> streamlines = [streamline, streamline[::2]]
>>> new_streamlines = set_number_of_points(streamlines, 10)
>>> [len(s) for s in streamlines]
[100, 50]
>>> [len(s) for s in new_streamlines]
[10, 10]


BundleMinDistanceAsymmetricMetric

class dipy.segment.bundles.BundleMinDistanceAsymmetricMetric(num_threads=None)

Asymmetric Bundle-based Minimum distance

This is a cost function that can be used by the StreamlineLinearRegistration class.

Methods

 distance(xopt) Distance calculated from this Metric setup(static, moving) Setup static and moving sets of streamlines
__init__(num_threads=None)

An abstract class for the metric used for streamline registration

If the two sets of streamlines match exactly then method distance of this object should be minimum.

Parameters

Number of threads. If None (default) then all available threads will be used. Only metrics using OpenMP will use this variable.

distance(xopt)

Distance calculated from this Metric

Parameters
xoptsequence

List of affine parameters as an 1D vector

BundleMinDistanceMetric

class dipy.segment.bundles.BundleMinDistanceMetric(num_threads=None)

Bundle-based Minimum Distance aka BMD

This is the cost function used by the StreamlineLinearRegistration

References

Garyfallidis14

Garyfallidis et al., “Direct native-space fiber bundle alignment for group comparisons”, ISMRM, 2014.

Methods

 setup(static, moving) distance(xopt)
__init__(num_threads=None)

An abstract class for the metric used for streamline registration

If the two sets of streamlines match exactly then method distance of this object should be minimum.

Parameters

Number of threads. If None (default) then all available threads will be used. Only metrics using OpenMP will use this variable.

distance(xopt)

Distance calculated from this Metric

Parameters
xoptsequence

List of affine parameters as an 1D vector,

setup(static, moving)

Setup static and moving sets of streamlines

Parameters
staticstreamlines

Fixed or reference set of streamlines.

movingstreamlines

Moving streamlines.

Number of threads. If None (default) then all available threads will be used.

Notes

Call this after the object is initiated and before distance.

BundleSumDistanceMatrixMetric

class dipy.segment.bundles.BundleSumDistanceMatrixMetric(num_threads=None)

Bundle-based Sum Distance aka BMD

This is a cost function that can be used by the StreamlineLinearRegistration class.

Notes

The difference with BundleMinDistanceMatrixMetric is that it uses uses the sum of the distance matrix and not the sum of mins.

Methods

 setup(static, moving) distance(xopt)
__init__(num_threads=None)

An abstract class for the metric used for streamline registration

If the two sets of streamlines match exactly then method distance of this object should be minimum.

Parameters

Number of threads. If None (default) then all available threads will be used. Only metrics using OpenMP will use this variable.

distance(xopt)

Distance calculated from this Metric

Parameters
xoptsequence

List of affine parameters as an 1D vector

RecoBundles

class dipy.segment.bundles.RecoBundles(streamlines, greater_than=50, less_than=1000000, cluster_map=None, clust_thr=15, nb_pts=20, rng=None, verbose=False)

Bases: object

Methods

 evaluate_results(model_bundle, …) Compare the similiarity between two given bundles, model bundle, and extracted bundle. recognize(model_bundle, model_clust_thr[, …]) Recognize the model_bundle in self.streamlines refine(model_bundle, pruned_streamlines, …) Refine and recognize the model_bundle in self.streamlines This method expects once pruned streamlines as input.
__init__(streamlines, greater_than=50, less_than=1000000, cluster_map=None, clust_thr=15, nb_pts=20, rng=None, verbose=False)

Recognition of bundles

Extract bundles from a participants’ tractograms using model bundles segmented from a different subject or an atlas of bundles. See [Garyfallidis17] for the details.

Parameters
streamlinesStreamlines

The tractogram in which you want to recognize bundles.

greater_thanint, optional

Keep streamlines that have length greater than this value (default 50)

less_thanint, optional

Keep streamlines have length less than this value (default 1000000)

cluster_mapQB map, optional.

Provide existing clustering to start RB faster (default None).

clust_thrfloat, optional.

Distance threshold in mm for clustering streamlines. Default: 15.

nb_ptsint, optional.

Number of points per streamline (default 20)

rngRandomState

If None define RandomState in initialization function. Default: None

verbose: bool, optional.

Notes

Make sure that before creating this class that the streamlines and the model bundles are roughly in the same space. Also default thresholds are assumed in RAS 1mm^3 space. You may want to adjust those if your streamlines are not in world coordinates.

References

Garyfallidis17

Garyfallidis et al. Recognition of white matter bundles using local and global streamline-based registration and clustering, Neuroimage, 2017.

evaluate_results(model_bundle, pruned_streamlines, slr_select)

Compare the similiarity between two given bundles, model bundle, and extracted bundle.

Parameters
model_bundleStreamlines
pruned_streamlinesStreamlines
slr_selecttuple

Select the number of streamlines from model to neirborhood of model to perform the local SLR.

Returns
ba_valuefloat

bundle adjacency value between model bundle and pruned bundle

bmd_valuefloat

bundle minimum distance value between model bundle and pruned bundle

recognize(model_bundle, model_clust_thr, reduction_thr=10, reduction_distance='mdf', slr=True, slr_num_threads=None, slr_metric=None, slr_x0=None, slr_bounds=None, slr_select=(400, 600), slr_method='L-BFGS-B', pruning_thr=5, pruning_distance='mdf')

Recognize the model_bundle in self.streamlines

Parameters
model_bundleStreamlines
model_clust_thrfloat
reduction_thrfloat
reduction_distancestring

mdf or mam (default mdf)

slrbool

Use Streamline-based Linear Registration (SLR) locally (default True)

slr_metricBundleMinDistanceMetric
slr_x0array

(default None)

slr_boundsarray

(default None)

slr_selecttuple

Select the number of streamlines from model to neirborhood of model to perform the local SLR.

slr_methodstring

Optimization method (default ‘L-BFGS-B’)

pruning_thrfloat
pruning_distancestring

MDF (‘mdf’) and MAM (‘mam’)

Returns
recognized_transfStreamlines

Recognized bundle in the space of the model tractogram

recognized_labelsarray

Indices of recognized bundle in the original tractogram

References

Garyfallidis17

Garyfallidis et al. Recognition of white matter bundles using local and global streamline-based registration and clustering, Neuroimage, 2017.

refine(model_bundle, pruned_streamlines, model_clust_thr, reduction_thr=14, reduction_distance='mdf', slr=True, slr_metric=None, slr_x0=None, slr_bounds=None, slr_select=(400, 600), slr_method='L-BFGS-B', pruning_thr=6, pruning_distance='mdf')

Refine and recognize the model_bundle in self.streamlines This method expects once pruned streamlines as input. It refines the first ouput of recobundle by applying second local slr (optional), and second pruning. This method is useful when we are dealing with noisy data or when we want to extract small tracks from tractograms.

Parameters
model_bundleStreamlines
pruned_streamlinesStreamlines
model_clust_thrfloat
reduction_thrfloat
reduction_distancestring

mdf or mam (default mam)

slrbool

Use Streamline-based Linear Registration (SLR) locally (default True)

slr_metricBundleMinDistanceMetric
slr_x0array

(default None)

slr_boundsarray

(default None)

slr_selecttuple

Select the number of streamlines from model to neirborhood of model to perform the local SLR.

slr_methodstring

Optimization method (default ‘L-BFGS-B’)

pruning_thrfloat
pruning_distancestring

MDF (‘mdf’) and MAM (‘mam’)

Returns
recognized_transfStreamlines

Recognized bundle in the space of the model tractogram

recognized_labelsarray

Indices of recognized bundle in the original tractogram

References

Garyfallidis17

Garyfallidis et al. Recognition of white matter bundles using local and global streamline-based registration and clustering, Neuroimage, 2017.

Chandio2020

Chandio, B.Q., Risacher, S.L., Pestilli, F.,

Bullock, D., Yeh, FC., Koudoro, S., Rokem, A., Harezlack, J., and Garyfallidis, E. Bundle analytics, a computational framework for investigating the shapes and profiles of brain pathways across populations. Sci Rep 10, 17149 (2020)

StreamlineLinearRegistration

class dipy.segment.bundles.StreamlineLinearRegistration(metric=None, x0='rigid', method='L-BFGS-B', bounds=None, verbose=False, options=None, evolution=False, num_threads=None)

Bases: object

Methods

 optimize(static, moving[, mat]) Find the minimum of the provided metric.
__init__(metric=None, x0='rigid', method='L-BFGS-B', bounds=None, verbose=False, options=None, evolution=False, num_threads=None)

Linear registration of 2 sets of streamlines [Garyfallidis15].

Parameters
metricStreamlineDistanceMetric,

If None and fast is False then the BMD distance is used. If fast is True then a faster implementation of BMD is used. Otherwise, use the given distance metric.

x0array or int or str

Initial parametrization for the optimization.

If 1D array with:

a) 6 elements then only rigid registration is performed with the 3 first elements for translation and 3 for rotation. b) 7 elements also isotropic scaling is performed (similarity). c) 12 elements then translation, rotation (in degrees), scaling and shearing is performed (affine).

Here is an example of x0 with 12 elements: x0=np.array([0, 10, 0, 40, 0, 0, 2., 1.5, 1, 0.1, -0.5, 0])

This has translation (0, 10, 0), rotation (40, 0, 0) in degrees, scaling (2., 1.5, 1) and shearing (0.1, -0.5, 0).

If int:
1. 6

x0 = np.array([0, 0, 0, 0, 0, 0])

2. 7

x0 = np.array([0, 0, 0, 0, 0, 0, 1.])

3. 12

x0 = np.array([0, 0, 0, 0, 0, 0, 1., 1., 1, 0, 0, 0])

If str:
1. “rigid”

x0 = np.array([0, 0, 0, 0, 0, 0])

2. “similarity”

x0 = np.array([0, 0, 0, 0, 0, 0, 1.])

3. “affine”

x0 = np.array([0, 0, 0, 0, 0, 0, 1., 1., 1, 0, 0, 0])

methodstr,

‘L_BFGS_B’ or ‘Powell’ optimizers can be used. Default is ‘L_BFGS_B’.

boundslist of tuples or None,

If method == ‘L_BFGS_B’ then we can use bounded optimization. For example for the six parameters of rigid rotation we can set the bounds = [(-30, 30), (-30, 30), (-30, 30),

(-45, 45), (-45, 45), (-45, 45)]

That means that we have set the bounds for the three translations and three rotation axes (in degrees).

verbosebool, optional.

If True, if True then information about the optimization is shown. Default: False.

optionsNone or dict,

Extra options to be used with the selected method.

evolutionboolean

If True save the transformation for each iteration of the optimizer. Default is False. Supported only with Scipy >= 0.11.

Number of threads. If None (default) then all available threads will be used. Only metrics using OpenMP will use this variable.

References

Garyfallidis15

Garyfallidis et al. “Robust and efficient linear registration of white-matter fascicles in the space of streamlines”, NeuroImage, 117, 124–140, 2015

Garyfallidis14

Garyfallidis et al., “Direct native-space fiber bundle alignment for group comparisons”, ISMRM, 2014.

Garyfallidis17

Garyfallidis et al. Recognition of white matter bundles using local and global streamline-based registration and clustering, Neuroimage, 2017.

optimize(static, moving, mat=None)

Find the minimum of the provided metric.

Parameters
staticstreamlines

Reference or fixed set of streamlines.

movingstreamlines

Moving set of streamlines.

matarray

Transformation (4, 4) matrix to start the registration. mat is applied to moving. Default value None which means that initial transformation will be generated by shifting the centers of moving and static sets of streamlines to the origin.

Returns
mapStreamlineRegistrationMap

Streamlines

dipy.segment.bundles.Streamlines

alias of nibabel.streamlines.array_sequence.ArraySequence

chain

class dipy.segment.bundles.chain

Bases: object

chain(*iterables) –> chain object

Return a chain object whose .__next__() method returns elements from the first iterable until it is exhausted, then elements from the next iterable, until all of the iterables are exhausted.

Methods

 from_iterable chain.from_iterable(iterable) –> chain object
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

from_iterable()

chain.from_iterable(iterable) –> chain object

Alternate chain() constructor taking a single iterable argument that evaluates lazily.

apply_affine

dipy.segment.bundles.apply_affine(aff, pts)

Apply affine matrix aff to points pts

Returns result of application of aff to the right of pts. The coordinate dimension of pts should be the last.

For the 3D case, aff will be shape (4,4) and pts will have final axis length 3 - maybe it will just be N by 3. The return value is the transformed points, in this case:

res = np.dot(aff[:3,:3], pts.T) + aff[:3,3:4]
transformed_pts = res.T


This routine is more general than 3D, in that aff can have any shape (N,N), and pts can have any shape, as long as the last dimension is for the coordinates, and is therefore length N-1.

Parameters
aff(N, N) array-like

Homogenous affine, for 3D points, will be 4 by 4. Contrary to first appearance, the affine will be applied on the left of pts.

pts(…, N-1) array-like

Points, where the last dimension contains the coordinates of each point. For 3D, the last dimension will be length 3.

Returns
transformed_pts(…, N-1) array

transformed points

Examples

>>> aff = np.array([[0,2,0,10],[3,0,0,11],[0,0,4,12],[0,0,0,1]])
>>> pts = np.array([[1,2,3],[2,3,4],[4,5,6],[6,7,8]])
>>> apply_affine(aff, pts)
array([[14, 14, 24],
[16, 17, 28],
[20, 23, 36],
[24, 29, 44]]...)


Just to show that in the simple 3D case, it is equivalent to:

>>> (np.dot(aff[:3,:3], pts.T) + aff[:3,3:4]).T
array([[14, 14, 24],
[16, 17, 28],
[20, 23, 36],
[24, 29, 44]]...)


But pts can be a more complicated shape:

>>> pts = pts.reshape((2,2,3))
>>> apply_affine(aff, pts)
array([[[14, 14, 24],
[16, 17, 28]],

[[20, 23, 36],
[24, 29, 44]]]...)


ba_analysis

dipy.segment.bundles.ba_analysis(recognized_bundle, expert_bundle, nb_pts=20, threshold=6.0)

Calculates bundle adjacency score between two given bundles

Parameters
recognized_bundleStreamlines

Extracted bundle from the whole brain tractogram (eg: AF_L)

expert_bundleStreamlines

Model bundle used as reference while extracting similar type bundle from inout tractogram

nb_ptsinteger (default 20)

Discretizing streamlines to have nb_pts number of points

thresholdfloat (default 6)

Threshold used for in computing bundle adjacency. Threshold controls how much strictness user wants while calculating bundle adjacency between two bundles. Smaller threshold means bundles should be strictly adjacent to get higher BA score.

Returns
Bundle adjacency score between two tracts

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

dipy.segment.bundles.bundle_adjacency(dtracks0, dtracks1, threshold)

Find bundle adjacency between two given tracks/bundles

Parameters
dtracks0Streamlines

White matter tract from one subject

dtracks1Streamlines

White matter tract from another subject

thresholdfloat

Threshold controls how much strictness user wants while calculating bundle adjacency between two bundles. Smaller threshold means bundles should be strictly adjacent to get higher BA score.

Returns
resFloat

Bundle adjacency score between two tracts

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

bundle_shape_similarity

dipy.segment.bundles.bundle_shape_similarity(bundle1, bundle2, rng, clust_thr=[5, 3, 1.5], threshold=6)

Calculates bundle shape similarity between two given bundles using bundle adjacency (BA) metric

Parameters
bundle1Streamlines

White matter tract from one subject (eg: AF_L)

bundle2Streamlines

White matter tract from another subject (eg: AF_L)

rngRandomState
clust_thrlist of float (default [5, 3, 1.5])

list of clustering thresholds used in quickbundlesX

thresholdfloat (default 6)

Threshold used for in computing bundle adjacency. Threshold controls how much strictness user wants while calculating shape similarity between two bundles. Smaller threshold means bundles should be strictly similar to get higher shape similarity score.

Returns
ba_valueFloat

Bundle similarity score between two tracts

References

Chandio2020

Chandio, B.Q., Risacher, S.L., Pestilli, F., Bullock, D.,

Yeh, FC., Koudoro, S., Rokem, A., Harezlack, J., and Garyfallidis, E. Bundle analytics, a computational framework for investigating the shapes and profiles of brain pathways across populations. Sci Rep 10, 17149 (2020)

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

bundles_distances_mam

dipy.segment.bundles.bundles_distances_mam()

Calculate distances between list of tracks A and list of tracks B

Parameters
tracksAsequence

of tracks as arrays, shape (N1,3) .. (Nm,3)

tracksBsequence

of tracks as arrays, shape (N1,3) .. (Nm,3)

metricstr

‘avg’, ‘min’, ‘max’

Returns
DMarray, shape (len(tracksA), len(tracksB))

distances between tracksA and tracksB according to metric

bundles_distances_mdf

dipy.segment.bundles.bundles_distances_mdf()

Calculate distances between list of tracks A and list of tracks B

All tracks need to have the same number of points

Parameters
tracksAsequence

of tracks as arrays, [(N,3) .. (N,3)]

tracksBsequence

of tracks as arrays, [(N,3) .. (N,3)]

Returns
DMarray, shape (len(tracksA), len(tracksB))

distances between tracksA and tracksB according to metric

check_range

dipy.segment.bundles.check_range(streamline, gt, lt)

cluster_bundle

dipy.segment.bundles.cluster_bundle(bundle, clust_thr, rng, nb_pts=20, select_randomly=500000)

Clusters bundles

Parameters
bundleStreamlines

White matter tract

clust_thrfloat

clustering threshold used in quickbundlesX

rngRandomState
nb_pts: integer (default 20)

Discretizing streamlines to have nb_points number of points

select_randomly: integer (default 500000)

Randomly select streamlines from the input bundle

Returns
centroidsStreamlines

clustered centroids of the input bundle

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

length

dipy.segment.bundles.length()

Euclidean length of streamlines

Length is in mm only if streamlines are expressed in world coordinates.

Parameters
streamlinesndarray or a list or dipy.tracking.Streamlines

If ndarray, must have shape (N,3) where N is the number of points of the streamline. If list, each item must be ndarray shape (Ni,3) where Ni is the number of points of streamline i. If dipy.tracking.Streamlines, its common_shape must be 3.

Returns
lengthsscalar or ndarray shape (N,)

If there is only one streamline, a scalar representing the length of the streamline. If there are several streamlines, ndarray containing the length of every streamline.

Examples

>>> from dipy.tracking.streamline import length
>>> import numpy as np
>>> streamline = np.array([[1, 1, 1], [2, 3, 4], [0, 0, 0]])
>>> expected_length = np.sqrt([1+2**2+3**2, 2**2+3**2+4**2]).sum()
>>> length(streamline) == expected_length
True
>>> streamlines = [streamline, np.vstack([streamline, streamline[::-1]])]
>>> expected_lengths = [expected_length, 2*expected_length]
>>> lengths = [length(streamlines[0]), length(streamlines[1])]
>>> np.allclose(lengths, expected_lengths)
True
>>> length([])
0.0
>>> length(np.array([[1, 2, 3]]))
0.0


nbytes

dipy.segment.bundles.nbytes(streamlines)

qbx_and_merge

dipy.segment.bundles.qbx_and_merge(streamlines, thresholds, nb_pts=20, select_randomly=None, rng=None, verbose=False)

Run QuickBundlesX and then run again on the centroids of the last layer

Running again QuickBundles at a layer has the effect of merging some of the clusters that maybe originally devided because of branching. This function help obtain a result at a QuickBundles quality but with QuickBundlesX speed. The merging phase has low cost because it is applied only on the centroids rather than the entire dataset.

Parameters
streamlinesStreamlines
thresholdssequence

List of distance thresholds for QuickBundlesX.

nb_ptsint

Number of points for discretizing each streamline

select_randomlyint

Randomly select a specific number of streamlines. If None all the streamlines are used.

rngRandomState

If None then RandomState is initialized internally.

verbosebool, optional.

Returns
——-
clustersobj

Contains the clusters of the last layer of QuickBundlesX after merging.

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

Garyfallidis16

Garyfallidis E. et al. QuickBundlesX: Sequential clustering of millions of streamlines in multiple levels of detail at record execution time. Proceedings of the, International Society of Magnetic Resonance in Medicine (ISMRM). Singapore, 4187, 2016.

select_random_set_of_streamlines

dipy.segment.bundles.select_random_set_of_streamlines(streamlines, select, rng=None)

Select a random set of streamlines

Parameters
streamlinesSteamlines

Object of 2D ndarrays of shape[-1]==3

selectint

Number of streamlines to select. If there are less streamlines than select then select=len(streamlines).

rngRandomState

Default None.

Returns
selected_streamlineslist

Notes

The same streamline will not be selected twice.

set_number_of_points

dipy.segment.bundles.set_number_of_points()
Change the number of points of streamlines

(either by downsampling or upsampling)

Change the number of points of streamlines in order to obtain nb_points-1 segments of equal length. Points of streamlines will be modified along the curve.

Parameters
streamlinesndarray or a list or dipy.tracking.Streamlines

If ndarray, must have shape (N,3) where N is the number of points of the streamline. If list, each item must be ndarray shape (Ni,3) where Ni is the number of points of streamline i. If dipy.tracking.Streamlines, its common_shape must be 3.

nb_pointsint

integer representing number of points wanted along the curve.

Returns
new_streamlinesndarray or a list or dipy.tracking.Streamlines

Results of the downsampling or upsampling process.

Examples

>>> from dipy.tracking.streamline import set_number_of_points
>>> import numpy as np


One streamline, a semi-circle:

>>> theta = np.pi*np.linspace(0, 1, 100)
>>> x = np.cos(theta)
>>> y = np.sin(theta)
>>> z = 0 * x
>>> streamline = np.vstack((x, y, z)).T
>>> modified_streamline = set_number_of_points(streamline, 3)
>>> len(modified_streamline)
3


Multiple streamlines:

>>> streamlines = [streamline, streamline[::2]]
>>> new_streamlines = set_number_of_points(streamlines, 10)
>>> [len(s) for s in streamlines]
[100, 50]
>>> [len(s) for s in new_streamlines]
[10, 10]


time

dipy.segment.bundles.time() → floating point number

Return the current time in seconds since the Epoch. Fractions of a second may be present if the system clock provides them.

ABCMeta

class dipy.segment.clustering.ABCMeta

Bases: type

Metaclass for defining Abstract Base Classes (ABCs).

Use this metaclass to create an ABC. An ABC can be subclassed directly, and then acts as a mix-in class. You can also register unrelated concrete classes (even built-in classes) and unrelated ABCs as ‘virtual subclasses’ – these and their descendants will be considered subclasses of the registering ABC by the built-in issubclass() function, but the registering ABC won’t show up in their MRO (Method Resolution Order) nor will method implementations defined by the registering ABC be callable (not even via super()).

Methods

 __call__(*args, **kwargs) Call self as a function. mro(/) Return a type’s method resolution order. register(subclass) Register a virtual subclass of an ABC.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

register(subclass)

Register a virtual subclass of an ABC.

Returns the subclass, to allow usage as a class decorator.

AveragePointwiseEuclideanMetric

class dipy.segment.clustering.AveragePointwiseEuclideanMetric

Bases: dipy.segment.metricspeed.SumPointwiseEuclideanMetric

Computes the average of pointwise Euclidean distances between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between the features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

The distance between two 2D sequential data:

s1       s2

0*   a    *0
\       |
\      |
1*     |
|  b  *1
|      \
2*      \
c    *2


is equal to $$(a+b+c)/3$$ where $$a$$ is the Euclidean distance between s1[0] and s2[0], $$b$$ between s1[1] and s2[1] and $$c$$ between s1[2] and s2[2].

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

Cluster

class dipy.segment.clustering.Cluster(id=0, indices=None, refdata=<dipy.segment.clustering.Identity object>)

Bases: object

Provides functionalities for interacting with a cluster.

Useful container to retrieve index of elements grouped together. If a reference to the data is provided to cluster_map, elements will be returned instead of their index when possible.

Parameters
cluster_mapClusterMap object

Reference to the set of clusters this cluster is being part of.

idint

Id of this cluster in its associated cluster_map object.

refdatalist (optional)

Actual elements that clustered indices refer to.

Notes

A cluster does not contain actual data but instead knows how to retrieve them using its ClusterMap object.

Methods

 assign(*indices) Assigns indices to this cluster.
__init__(id=0, indices=None, refdata=<dipy.segment.clustering.Identity object>)

Initialize self. See help(type(self)) for accurate signature.

assign(*indices)

Assigns indices to this cluster.

Parameters
*indiceslist of indices

Indices to add to this cluster.

ClusterCentroid

class dipy.segment.clustering.ClusterCentroid(centroid, id=0, indices=None, refdata=<dipy.segment.clustering.Identity object>)

Provides functionalities for interacting with a cluster.

Useful container to retrieve the indices of elements grouped together and the cluster’s centroid. If a reference to the data is provided to cluster_map, elements will be returned instead of their index when possible.

Parameters
cluster_mapClusterMapCentroid object

Reference to the set of clusters this cluster is being part of.

idint

Id of this cluster in its associated cluster_map object.

refdatalist (optional)

Actual elements that clustered indices refer to.

Notes

A cluster does not contain actual data but instead knows how to retrieve them using its ClusterMapCentroid object.

Methods

 assign(id_datum, features) Assigns a data point to this cluster. Update centroid of this cluster.
__init__(centroid, id=0, indices=None, refdata=<dipy.segment.clustering.Identity object>)

Initialize self. See help(type(self)) for accurate signature.

assign(id_datum, features)

Assigns a data point to this cluster.

Parameters
id_datumint

Index of the data point to add to this cluster.

features2D array

Data point’s features to modify this cluster’s centroid.

update()

Update centroid of this cluster.

Returns
convergedbool

Tells if the centroid has moved.

ClusterMap

class dipy.segment.clustering.ClusterMap(refdata=<dipy.segment.clustering.Identity object>)

Bases: object

Provides functionalities for interacting with clustering outputs.

Useful container to create, remove, retrieve and filter clusters. If refdata is given, elements will be returned instead of their index when using Cluster objects.

Parameters
refdatalist

Actual elements that clustered indices refer to.

Attributes
clusters
refdata

Methods

 add_cluster(*clusters) Adds one or multiple clusters to this cluster map. Remove all clusters from this cluster map. Gets the size of every cluster contained in this cluster map. get_large_clusters(min_size) Gets clusters which contains at least min_size elements. get_small_clusters(max_size) Gets clusters which contains at most max_size elements. remove_cluster(*clusters) Remove one or multiple clusters from this cluster map. Gets number of clusters contained in this cluster map.
__init__(refdata=<dipy.segment.clustering.Identity object>)

Initialize self. See help(type(self)) for accurate signature.

add_cluster(*clusters)

Adds one or multiple clusters to this cluster map.

Parameters
*clustersCluster object, …

Cluster(s) to be added in this cluster map.

clear()

Remove all clusters from this cluster map.

property clusters
clusters_sizes()

Gets the size of every cluster contained in this cluster map.

Returns
list of int

Sizes of every cluster in this cluster map.

get_large_clusters(min_size)

Gets clusters which contains at least min_size elements.

Parameters
min_sizeint

Minimum number of elements a cluster needs to have to be selected.

Returns
list of Cluster objects

Clusters having at least min_size elements.

get_small_clusters(max_size)

Gets clusters which contains at most max_size elements.

Parameters
max_sizeint

Maximum number of elements a cluster can have to be selected.

Returns
list of Cluster objects

Clusters having at most max_size elements.

property refdata
remove_cluster(*clusters)

Remove one or multiple clusters from this cluster map.

Parameters
*clustersCluster object, …

Cluster(s) to be removed from this cluster map.

size()

Gets number of clusters contained in this cluster map.

ClusterMapCentroid

class dipy.segment.clustering.ClusterMapCentroid(refdata=<dipy.segment.clustering.Identity object>)

Provides functionalities for interacting with clustering outputs that have centroids.

Allows to retrieve easely the centroid of every cluster. Also, it is a useful container to create, remove, retrieve and filter clusters. If refdata is given, elements will be returned instead of their index when using ClusterCentroid objects.

Parameters
refdatalist

Actual elements that clustered indices refer to.

Attributes
centroids
clusters
refdata

Methods

 add_cluster(*clusters) Adds one or multiple clusters to this cluster map. clear() Remove all clusters from this cluster map. clusters_sizes() Gets the size of every cluster contained in this cluster map. get_large_clusters(min_size) Gets clusters which contains at least min_size elements. get_small_clusters(max_size) Gets clusters which contains at most max_size elements. remove_cluster(*clusters) Remove one or multiple clusters from this cluster map. size() Gets number of clusters contained in this cluster map.
__init__(refdata=<dipy.segment.clustering.Identity object>)

Initialize self. See help(type(self)) for accurate signature.

property centroids

Clustering

class dipy.segment.clustering.Clustering

Bases: object

Methods

 cluster(data[, ordering]) Clusters data.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

abstract cluster(data, ordering=None)

Clusters data.

Subclasses will perform their clustering algorithm here.

Parameters
datalist of N-dimensional arrays

Each array represents a data point.

orderingiterable of indices, optional

Specifies the order in which data points will be clustered.

Returns
ClusterMap object

Result of the clustering.

Identity

class dipy.segment.clustering.Identity

Bases: object

Provides identity indexing functionality.

This can replace any class supporting indexing used for referencing (e.g. list, tuple). Indexing an instance of this class will return the index provided instead of the element. It does not support slicing.

__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

Metric

class dipy.segment.clustering.Metric

Bases: object

Computes a distance between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between extracted features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

When subclassing Metric, one only needs to override the dist and are_compatible methods.

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

are_compatible()

Checks if features can be used by metric.dist based on their shape.

Basically this method exists so we don’t have to do this check inside the metric.dist function (speedup).

Parameters
shape1int, 1-tuple or 2-tuple

shape of the first data point’s features

shape2int, 1-tuple or 2-tuple

shape of the second data point’s features

Returns
are_compatiblebool

whether or not shapes are compatible

dist()

Computes a distance between two data points based on their features.

Parameters
features12D array

Features of the first data point.

features22D array

Features of the second data point.

Returns
double

Distance between two data points.

feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

MinimumAverageDirectFlipMetric

class dipy.segment.clustering.MinimumAverageDirectFlipMetric

Bases: dipy.segment.metricspeed.AveragePointwiseEuclideanMetric

Computes the MDF distance (minimum average direct-flip) between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

Notes

The distance between two 2D sequential data:

s1       s2

0*   a    *0
\       |
\      |
1*     |
|  b  *1
|      \
2*      \
c    *2


is equal to $$\min((a+b+c)/3, (a'+b'+c')/3)$$ where $$a$$ is the Euclidean distance between s1[0] and s2[0], $$b$$ between s1[1] and s2[1], $$c$$ between s1[2] and s2[2], $$a'$$ between s1[0] and s2[2], $$b'$$ between s1[1] and s2[1] and $$c'$$ between s1[2] and s2[0].

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

is_order_invariant

Is this metric invariant to the sequence’s ordering

QuickBundles

class dipy.segment.clustering.QuickBundles(threshold, metric='MDF_12points', max_nb_clusters=2147483647)

Clusters streamlines using QuickBundles [Garyfallidis12].

Given a list of streamlines, the QuickBundles algorithm sequentially assigns each streamline to its closest bundle in $$\mathcal{O}(Nk)$$ where $$N$$ is the number of streamlines and $$k$$ is the final number of bundles. If for a given streamline its closest bundle is farther than threshold, a new bundle is created and the streamline is assigned to it except if the number of bundles has already exceeded max_nb_clusters.

Parameters
thresholdfloat

The maximum distance from a bundle for a streamline to be still considered as part of it.

metricstr or Metric object (optional)

The distance metric to use when comparing two streamlines. By default, the Minimum average Direct-Flip (MDF) distance [Garyfallidis12] is used and streamlines are automatically resampled so they have 12 points.

max_nb_clustersint

Limits the creation of bundles.

References

Garyfallidis12(1,2,3)

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

Examples

>>> from dipy.segment.clustering import QuickBundles
>>> from dipy.data import get_fnames
>>> from dipy.tracking.streamline import Streamlines
>>> fname = get_fnames('fornix')
...                          bbox_valid_check=False).streamlines
>>> streamlines = Streamlines(fornix)
>>> # Segment fornix with a threshold of 10mm and streamlines resampled
>>> # to 12 points.
>>> qb = QuickBundles(threshold=10.)
>>> clusters = qb.cluster(streamlines)
>>> len(clusters)
4
>>> list(map(len, clusters))
[61, 191, 47, 1]
>>> # Resampling streamlines differently is done explicitly as follows.
>>> # Note this has an impact on the speed and the accuracy (tradeoff).
>>> from dipy.segment.metric import ResampleFeature
>>> from dipy.segment.metric import AveragePointwiseEuclideanMetric
>>> feature = ResampleFeature(nb_points=2)
>>> metric = AveragePointwiseEuclideanMetric(feature)
>>> qb = QuickBundles(threshold=10., metric=metric)
>>> clusters = qb.cluster(streamlines)
>>> len(clusters)
4
>>> list(map(len, clusters))
[58, 142, 72, 28]


Methods

 cluster(streamlines[, ordering]) Clusters streamlines into bundles.
__init__(threshold, metric='MDF_12points', max_nb_clusters=2147483647)

Initialize self. See help(type(self)) for accurate signature.

cluster(streamlines, ordering=None)

Clusters streamlines into bundles.

Performs quickbundles algorithm using predefined metric and threshold.

Parameters
streamlineslist of 2D arrays

Each 2D array represents a sequence of 3D points (points, 3).

orderingiterable of indices

Specifies the order in which data points will be clustered.

Returns
ClusterMapCentroid object

Result of the clustering.

QuickBundlesX

class dipy.segment.clustering.QuickBundlesX(thresholds, metric='MDF_12points')

Clusters streamlines using QuickBundlesX.

Parameters
thresholdslist of float

Thresholds to use for each clustering layer. A threshold represents the maximum distance from a cluster for a streamline to be still considered as part of it.

metricstr or Metric object (optional)

The distance metric to use when comparing two streamlines. By default, the Minimum average Direct-Flip (MDF) distance [Garyfallidis12] is used and streamlines are automatically resampled so they have 12 points.

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

Garyfallidis16

Garyfallidis E. et al. QuickBundlesX: Sequential clustering of millions of streamlines in multiple levels of detail at record execution time. Proceedings of the, International Society of Magnetic Resonance in Medicine (ISMRM). Singapore, 4187, 2016.

Methods

 cluster(streamlines[, ordering]) Clusters streamlines into bundles.
__init__(thresholds, metric='MDF_12points')

Initialize self. See help(type(self)) for accurate signature.

cluster(streamlines, ordering=None)

Clusters streamlines into bundles.

Performs QuickbundleX using a predefined metric and thresholds.

Parameters
streamlineslist of 2D arrays

Each 2D array represents a sequence of 3D points (points, 3).

orderingiterable of indices

Specifies the order in which data points will be clustered.

Returns
TreeClusterMap object

Result of the clustering.

ResampleFeature

class dipy.segment.clustering.ResampleFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The features being extracted are the points of the sequence once resampled. This is useful for metrics requiring a constant number of points for all

streamlines.

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

TreeCluster

class dipy.segment.clustering.TreeCluster(threshold, centroid, indices=None)
Attributes
is_leaf

Methods

 assign(id_datum, features) Assigns a data point to this cluster. update() Update centroid of this cluster.
__init__(threshold, centroid, indices=None)

Initialize self. See help(type(self)) for accurate signature.

add(child)
property is_leaf
return_indices()

TreeClusterMap

class dipy.segment.clustering.TreeClusterMap(root)
Attributes
clusters
refdata

Methods

 add_cluster(*clusters) Adds one or multiple clusters to this cluster map. clear() Remove all clusters from this cluster map. clusters_sizes() Gets the size of every cluster contained in this cluster map. get_large_clusters(min_size) Gets clusters which contains at least min_size elements. get_small_clusters(max_size) Gets clusters which contains at most max_size elements. remove_cluster(*clusters) Remove one or multiple clusters from this cluster map. size() Gets number of clusters contained in this cluster map.
 get_clusters iter_preorder traverse_postorder
__init__(root)

Initialize self. See help(type(self)) for accurate signature.

get_clusters(wanted_level)
iter_preorder(node)
property refdata
traverse_postorder(node, visit)

abstractmethod

dipy.segment.clustering.abstractmethod(funcobj)

A decorator indicating abstract methods.

Requires that the metaclass is ABCMeta or derived from it. A class that has a metaclass derived from ABCMeta cannot be instantiated unless all of its abstract methods are overridden. The abstract methods can be called using any of the normal ‘super’ call mechanisms.

Usage:

class C(metaclass=ABCMeta):

@abstractmethod def my_abstract_method(self, …):

nbytes

dipy.segment.clustering.nbytes(streamlines)

qbx_and_merge

dipy.segment.clustering.qbx_and_merge(streamlines, thresholds, nb_pts=20, select_randomly=None, rng=None, verbose=False)

Run QuickBundlesX and then run again on the centroids of the last layer

Running again QuickBundles at a layer has the effect of merging some of the clusters that maybe originally devided because of branching. This function help obtain a result at a QuickBundles quality but with QuickBundlesX speed. The merging phase has low cost because it is applied only on the centroids rather than the entire dataset.

Parameters
streamlinesStreamlines
thresholdssequence

List of distance thresholds for QuickBundlesX.

nb_ptsint

Number of points for discretizing each streamline

select_randomlyint

Randomly select a specific number of streamlines. If None all the streamlines are used.

rngRandomState

If None then RandomState is initialized internally.

verbosebool, optional.

Returns
——-
clustersobj

Contains the clusters of the last layer of QuickBundlesX after merging.

References

Garyfallidis12

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

Garyfallidis16

Garyfallidis E. et al. QuickBundlesX: Sequential clustering of millions of streamlines in multiple levels of detail at record execution time. Proceedings of the, International Society of Magnetic Resonance in Medicine (ISMRM). Singapore, 4187, 2016.

set_number_of_points

dipy.segment.clustering.set_number_of_points()
Change the number of points of streamlines

(either by downsampling or upsampling)

Change the number of points of streamlines in order to obtain nb_points-1 segments of equal length. Points of streamlines will be modified along the curve.

Parameters
streamlinesndarray or a list or dipy.tracking.Streamlines

If ndarray, must have shape (N,3) where N is the number of points of the streamline. If list, each item must be ndarray shape (Ni,3) where Ni is the number of points of streamline i. If dipy.tracking.Streamlines, its common_shape must be 3.

nb_pointsint

integer representing number of points wanted along the curve.

Returns
new_streamlinesndarray or a list or dipy.tracking.Streamlines

Results of the downsampling or upsampling process.

Examples

>>> from dipy.tracking.streamline import set_number_of_points
>>> import numpy as np


One streamline, a semi-circle:

>>> theta = np.pi*np.linspace(0, 1, 100)
>>> x = np.cos(theta)
>>> y = np.sin(theta)
>>> z = 0 * x
>>> streamline = np.vstack((x, y, z)).T
>>> modified_streamline = set_number_of_points(streamline, 3)
>>> len(modified_streamline)
3


Multiple streamlines:

>>> streamlines = [streamline, streamline[::2]]
>>> new_streamlines = set_number_of_points(streamlines, 10)
>>> [len(s) for s in streamlines]
[100, 50]
>>> [len(s) for s in new_streamlines]
[10, 10]


time

dipy.segment.clustering.time() → floating point number

Return the current time in seconds since the Epoch. Fractions of a second may be present if the system clock provides them.

dipy.segment.mask.applymask(vol, mask)

Parameters
volndarray

Array with $$V$$ dimensions

Binary mask. Has $$M$$ dimensions where $$M <= V$$. When $$M < V$$, we append $$V - M$$ dimensions with axis length 1 to mask so that mask will broadcast against vol. In the typical case vol can be 4D, mask can be 3D, and we append a 1 to the mask shape which (via numpy broadcasting) has the effect of appling the 3D mask to each 3D slice in vol (vol[..., 0] to vol[..., -1).

Returns

vol multiplied by mask where mask may have been extended to match extra dimensions in vol

binary_dilation

dipy.segment.mask.binary_dilation(input, structure=None, iterations=1, mask=None, output=None, border_value=0, origin=0, brute_force=False)

Multidimensional binary dilation with the given structuring element.

Parameters
inputarray_like

Binary array_like to be dilated. Non-zero (True) elements form the subset to be dilated.

structurearray_like, optional

Structuring element used for the dilation. Non-zero elements are considered True. If no structuring element is provided an element is generated with a square connectivity equal to one.

iterationsint, optional

The dilation is repeated iterations times (one, by default). If iterations is less than 1, the dilation is repeated until the result does not change anymore. Only an integer of iterations is accepted.

If a mask is given, only those elements with a True value at the corresponding mask element are modified at each iteration.

outputndarray, optional

Array of the same shape as input, into which the output is placed. By default, a new array is created.

border_valueint (cast to 0 or 1), optional

Value at the border in the output array.

originint or tuple of ints, optional

Placement of the filter, by default 0.

brute_forceboolean, optional

Memory condition: if False, only the pixels whose value was changed in the last iteration are tracked as candidates to be updated (dilated) in the current iteration; if True all pixels are considered as candidates for dilation, regardless of what happened in the previous iteration. False by default.

Returns
binary_dilationndarray of bools

Dilation of the input by the structuring element.

grey_dilation, binary_erosion, binary_closing, binary_opening
generate_binary_structure

Notes

Dilation [1] is a mathematical morphology operation [2] that uses a structuring element for expanding the shapes in an image. The binary dilation of an image by a structuring element is the locus of the points covered by the structuring element, when its center lies within the non-zero points of the image.

References

1

https://en.wikipedia.org/wiki/Dilation_%28morphology%29

2

https://en.wikipedia.org/wiki/Mathematical_morphology

Examples

>>> from scipy import ndimage
>>> a = np.zeros((5, 5))
>>> a[2, 2] = 1
>>> a
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> ndimage.binary_dilation(a)
array([[False, False, False, False, False],
[False, False,  True, False, False],
[False,  True,  True,  True, False],
[False, False,  True, False, False],
[False, False, False, False, False]], dtype=bool)
>>> ndimage.binary_dilation(a).astype(a.dtype)
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> # 3x3 structuring element with connectivity 1, used by default
>>> struct1 = ndimage.generate_binary_structure(2, 1)
>>> struct1
array([[False,  True, False],
[ True,  True,  True],
[False,  True, False]], dtype=bool)
>>> # 3x3 structuring element with connectivity 2
>>> struct2 = ndimage.generate_binary_structure(2, 2)
>>> struct2
array([[ True,  True,  True],
[ True,  True,  True],
[ True,  True,  True]], dtype=bool)
>>> ndimage.binary_dilation(a, structure=struct1).astype(a.dtype)
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> ndimage.binary_dilation(a, structure=struct2).astype(a.dtype)
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> ndimage.binary_dilation(a, structure=struct1,\
... iterations=2).astype(a.dtype)
array([[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 1.,  1.,  1.,  1.,  1.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  1.,  0.,  0.]])


bounding_box

dipy.segment.mask.bounding_box(vol)

Compute the bounding box of nonzero intensity voxels in the volume.

Parameters
volndarray

Volume to compute bounding box on.

Returns
npminslist

Array containg minimum index of each dimension

npmaxslist

Array containg maximum index of each dimension

dipy.segment.mask.clean_cc_mask(mask)

Cleans a segmentation of the corpus callosum so no random pixels are included.

Parameters

Binary mask of the coarse segmentation.

Returns

Binary mask of the cleaned segmentation.

color_fa

dipy.segment.mask.color_fa(fa, evecs)

Color fractional anisotropy of diffusion tensor

Parameters
faarray-like

Array of the fractional anisotropy (can be 1D, 2D or 3D)

evecsarray-like

eigen vectors from the tensor model

Returns
rgbArray with 3 channels for each color as the last dimension.

Colormap of the FA with red for the x value, y for the green value and z for the blue value.

Notes

It is computed from the clipped FA between 0 and 1 using the following formula

$rgb = abs(max(\vec{e})) \times fa$

crop

dipy.segment.mask.crop(vol, mins, maxs)

Crops the input volume.

Parameters
volndarray

Volume to crop.

minsarray

Array containg minimum index of each dimension.

maxsarray

Array containg maximum index of each dimension.

Returns
volndarray

The cropped volume.

fractional_anisotropy

dipy.segment.mask.fractional_anisotropy(evals, axis=-1)

Return Fractional anisotropy (FA) of a diffusion tensor.

Parameters
evalsarray-like

Eigenvalues of a diffusion tensor.

axisint

Axis of evals which contains 3 eigenvalues.

Returns
faarray

Calculated FA. Range is 0 <= FA <= 1.

Notes

FA is calculated using the following equation:

$FA = \sqrt{\frac{1}{2}\frac{(\lambda_1-\lambda_2)^2+(\lambda_1- \lambda_3)^2+(\lambda_2-\lambda_3)^2}{\lambda_1^2+ \lambda_2^2+\lambda_3^2}}$

generate_binary_structure

dipy.segment.mask.generate_binary_structure(rank, connectivity)

Generate a binary structure for binary morphological operations.

Parameters
rankint

Number of dimensions of the array to which the structuring element will be applied, as returned by np.ndim.

connectivityint

connectivity determines which elements of the output array belong to the structure, i.e., are considered as neighbors of the central element. Elements up to a squared distance of connectivity from the center are considered neighbors. connectivity may range from 1 (no diagonal elements are neighbors) to rank (all elements are neighbors).

Returns
outputndarray of bools

Structuring element which may be used for binary morphological operations, with rank dimensions and all dimensions equal to 3.

iterate_structure, binary_dilation, binary_erosion

Notes

generate_binary_structure can only create structuring elements with dimensions equal to 3, i.e., minimal dimensions. For larger structuring elements, that are useful e.g., for eroding large objects, one may either use iterate_structure, or create directly custom arrays with numpy functions such as numpy.ones.

Examples

>>> from scipy import ndimage
>>> struct = ndimage.generate_binary_structure(2, 1)
>>> struct
array([[False,  True, False],
[ True,  True,  True],
[False,  True, False]], dtype=bool)
>>> a = np.zeros((5,5))
>>> a[2, 2] = 1
>>> a
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> b = ndimage.binary_dilation(a, structure=struct).astype(a.dtype)
>>> b
array([[ 0.,  0.,  0.,  0.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  0.,  0.,  0.,  0.]])
>>> ndimage.binary_dilation(b, structure=struct).astype(a.dtype)
array([[ 0.,  0.,  1.,  0.,  0.],
[ 0.,  1.,  1.,  1.,  0.],
[ 1.,  1.,  1.,  1.,  1.],
[ 0.,  1.,  1.,  1.,  0.],
[ 0.,  0.,  1.,  0.,  0.]])
>>> struct = ndimage.generate_binary_structure(2, 2)
>>> struct
array([[ True,  True,  True],
[ True,  True,  True],
[ True,  True,  True]], dtype=bool)
>>> struct = ndimage.generate_binary_structure(3, 1)
>>> struct # no diagonal elements
array([[[False, False, False],
[False,  True, False],
[False, False, False]],
[[False,  True, False],
[ True,  True,  True],
[False,  True, False]],
[[False, False, False],
[False,  True, False],
[False, False, False]]], dtype=bool)


median_filter

dipy.segment.mask.median_filter(input, size=None, footprint=None, output=None, mode='reflect', cval=0.0, origin=0)

Calculate a multidimensional median filter.

Parameters
inputarray_like

The input array.

sizescalar or tuple, optional

See footprint, below. Ignored if footprint is given.

footprintarray, optional

Either size or footprint must be defined. size gives the shape that is taken from the input array, at every element position, to define the input to the filter function. footprint is a boolean array that specifies (implicitly) a shape, but also which of the elements within this shape will get passed to the filter function. Thus size=(n,m) is equivalent to footprint=np.ones((n,m)). We adjust size to the number of dimensions of the input array, so that, if the input array is shape (10,10,10), and size is 2, then the actual size used is (2,2,2). When footprint is given, size is ignored.

outputarray or dtype, optional

The array in which to place the output, or the dtype of the returned array. By default an array of the same dtype as input will be created.

mode{‘reflect’, ‘constant’, ‘nearest’, ‘mirror’, ‘wrap’}, optional

The mode parameter determines how the input array is extended beyond its boundaries. Default is ‘reflect’. Behavior for each valid value is as follows:

‘reflect’ (d c b a | a b c d | d c b a)

The input is extended by reflecting about the edge of the last pixel.

‘constant’ (k k k k | a b c d | k k k k)

The input is extended by filling all values beyond the edge with the same constant value, defined by the cval parameter.

‘nearest’ (a a a a | a b c d | d d d d)

The input is extended by replicating the last pixel.

‘mirror’ (d c b | a b c d | c b a)

The input is extended by reflecting about the center of the last pixel.

‘wrap’ (a b c d | a b c d | a b c d)

The input is extended by wrapping around to the opposite edge.

cvalscalar, optional

Value to fill past edges of input if mode is ‘constant’. Default is 0.0.

originint or sequence, optional

Controls the placement of the filter on the input array’s pixels. A value of 0 (the default) centers the filter over the pixel, with positive values shifting the filter to the left, and negative ones to the right. By passing a sequence of origins with length equal to the number of dimensions of the input array, different shifts can be specified along each axis.

Returns
median_filterndarray

Filtered array. Has the same shape as input.

Examples

>>> from scipy import ndimage, misc
>>> import matplotlib.pyplot as plt
>>> fig = plt.figure()
>>> plt.gray()  # show the filtered result in grayscale
>>> ax1 = fig.add_subplot(121)  # left side
>>> ax2 = fig.add_subplot(122)  # right side
>>> ascent = misc.ascent()
>>> result = ndimage.median_filter(ascent, size=20)
>>> ax1.imshow(ascent)
>>> ax2.imshow(result)
>>> plt.show()


median_otsu

dipy.segment.mask.median_otsu(input_volume, vol_idx=None, median_radius=4, numpass=4, autocrop=False, dilate=None)

Simple brain extraction tool method for images from DWI data.

It uses a median filter smoothing of the input_volumes vol_idx and an automatic histogram Otsu thresholding technique, hence the name median_otsu.

This function is inspired from Mrtrix’s bet which has default values median_radius=3, numpass=2. However, from tests on multiple 1.5T and 3T data from GE, Philips, Siemens, the most robust choice is median_radius=4, numpass=4.

Parameters
input_volumendarray

3D or 4D array of the brain volume.

vol_idxNone or array, optional.

1D array representing indices of axis=3 of a 4D input_volume. None is only an acceptable input if input_volume is 3D.

Radius (in voxels) of the applied median filter (default: 4).

numpass: int

Number of pass of the median filter (default: 4).

autocrop: bool, optional

if True, the masked input_volume will also be cropped using the bounding box defined by the masked data. Should be on if DWI is upsampled to 1x1x1 resolution. (default: False).

dilateNone or int, optional

number of iterations for binary dilation

Returns

Notes

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

3. Neither the name of skimage nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE AUTHOR AS IS’’ AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

multi_median

dipy.segment.mask.multi_median(input, median_radius, numpass)

Applies median filter multiple times on input data.

Parameters
inputndarray

The input volume to apply filter on.

Radius (in voxels) of the applied median filter

numpass: int

Number of pass of the median filter

Returns
inputndarray

Filtered input volume.

otsu

dipy.segment.mask.otsu(image, nbins=256)

Return threshold value based on Otsu’s method.

Parameters
image(N, M) ndarray

Grayscale input image.

nbinsint, optional

Number of bins used to calculate histogram. This value is ignored for integer arrays.

Returns
thresholdfloat

Upper threshold value. All pixels with an intensity higher than this value are assumed to be foreground.

Raises
ValueError

If image only contains a single grayscale value.

Notes

The input image must be grayscale.

References

1

Examples

>>> from skimage.data import camera
>>> image = camera()
>>> thresh = threshold_otsu(image)
>>> binary = image <= thresh


segment_from_cfa

dipy.segment.mask.segment_from_cfa(tensor_fit, roi, threshold, return_cfa=False)

Segment the cfa inside roi using the values from threshold as bounds.

Parameters
tensor_fitTensorFit object

TensorFit object

roindarray

A binary mask, which contains the bounding box for the segmentation.

thresholdarray-like

An iterable that defines the min and max values to use for the thresholding. The values are specified as (R_min, R_max, G_min, G_max, B_min, B_max)

return_cfabool, optional

If True, the cfa is also returned.

Returns

cfandarray, optional

Array with shape = (…, 3), where … is the shape of tensor_fit. The color fractional anisotropy, ordered as a nd array with the last dimension of size 3 for the R, G and B channels.

warn

dipy.segment.mask.warn(/, message, category=None, stacklevel=1, source=None)

Issue a warning, or maybe ignore it or raise an exception.

ArcLengthFeature

class dipy.segment.metric.ArcLengthFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The feature being extracted consists of one scalar representing the arc length of the sequence (i.e. the sum of the length of all segments).

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

AveragePointwiseEuclideanMetric

class dipy.segment.metric.AveragePointwiseEuclideanMetric

Bases: dipy.segment.metricspeed.SumPointwiseEuclideanMetric

Computes the average of pointwise Euclidean distances between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between the features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

The distance between two 2D sequential data:

s1       s2

0*   a    *0
\       |
\      |
1*     |
|  b  *1
|      \
2*      \
c    *2


is equal to $$(a+b+c)/3$$ where $$a$$ is the Euclidean distance between s1[0] and s2[0], $$b$$ between s1[1] and s2[1] and $$c$$ between s1[2] and s2[2].

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

CenterOfMassFeature

class dipy.segment.metric.CenterOfMassFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The feature being extracted consists of one N-dimensional point representing the mean of the points, i.e. the center of mass.

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

CosineMetric

class dipy.segment.metric.CosineMetric

Bases: dipy.segment.metricspeed.CythonMetric

Computes the cosine distance between two vectors.

A vector (i.e. a N-dimensional point) is represented as a 2D array with shape (1, nb_dimensions).

Notes

The distance between two vectors $$v_1$$ and $$v_2$$ is equal to $$\frac{1}{\pi} \arccos\left(\frac{v_1 \cdot v_2}{\|v_1\| \|v_2\|}\right)$$ and is bounded within $$[0,1]$$.

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

EuclideanMetric

dipy.segment.metric.EuclideanMetric

alias of dipy.segment.metricspeed.SumPointwiseEuclideanMetric

Feature

class dipy.segment.metric.Feature

Bases: object

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

Parameters
is_order_invariantbool (optional)

tells if this feature is invariant to the sequence’s ordering. This means starting from either extremities produces the same features. (Default: True)

Notes

When subclassing Feature, one only needs to override the extract and infer_shape methods.

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

extract()

Extracts features from a sequential datum.

Parameters
datum2D array

Sequence of N-dimensional points.

Returns
2D array

Features extracted from datum.

infer_shape()

Infers the shape of features extracted from a sequential datum.

Parameters
datum2D array

Sequence of N-dimensional points.

Returns
int, 1-tuple or 2-tuple

Shape of the features.

is_order_invariant

Is this feature invariant to the sequence’s ordering

IdentityFeature

class dipy.segment.metric.IdentityFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The features being extracted are the actual sequence’s points. This is useful for metric that does not require any pre-processing.

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

Metric

class dipy.segment.metric.Metric

Bases: object

Computes a distance between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between extracted features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

When subclassing Metric, one only needs to override the dist and are_compatible methods.

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

are_compatible()

Checks if features can be used by metric.dist based on their shape.

Basically this method exists so we don’t have to do this check inside the metric.dist function (speedup).

Parameters
shape1int, 1-tuple or 2-tuple

shape of the first data point’s features

shape2int, 1-tuple or 2-tuple

shape of the second data point’s features

Returns
are_compatiblebool

whether or not shapes are compatible

dist()

Computes a distance between two data points based on their features.

Parameters
features12D array

Features of the first data point.

features22D array

Features of the second data point.

Returns
double

Distance between two data points.

feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

MidpointFeature

class dipy.segment.metric.MidpointFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The feature being extracted consists of one N-dimensional point representing the middle point of the sequence (i.e. nb_points//2th point).

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

MinimumAverageDirectFlipMetric

class dipy.segment.metric.MinimumAverageDirectFlipMetric

Bases: dipy.segment.metricspeed.AveragePointwiseEuclideanMetric

Computes the MDF distance (minimum average direct-flip) between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

Notes

The distance between two 2D sequential data:

s1       s2

0*   a    *0
\       |
\      |
1*     |
|  b  *1
|      \
2*      \
c    *2


is equal to $$\min((a+b+c)/3, (a'+b'+c')/3)$$ where $$a$$ is the Euclidean distance between s1[0] and s2[0], $$b$$ between s1[1] and s2[1], $$c$$ between s1[2] and s2[2], $$a'$$ between s1[0] and s2[2], $$b'$$ between s1[1] and s2[1] and $$c'$$ between s1[2] and s2[0].

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

is_order_invariant

Is this metric invariant to the sequence’s ordering

ResampleFeature

class dipy.segment.metric.ResampleFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The features being extracted are the points of the sequence once resampled. This is useful for metrics requiring a constant number of points for all

streamlines.

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

SumPointwiseEuclideanMetric

class dipy.segment.metric.SumPointwiseEuclideanMetric

Bases: dipy.segment.metricspeed.CythonMetric

Computes the sum of pointwise Euclidean distances between two sequential data.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions). A feature object can be specified in order to calculate the distance between the features, rather than directly between the sequential data.

Parameters
featureFeature object, optional

It is used to extract features before computing the distance.

Notes

The distance between two 2D sequential data:

s1       s2

0*   a    *0
\       |
\      |
1*     |
|  b  *1
|      \
2*      \
c    *2


is equal to $$a+b+c$$ where $$a$$ is the Euclidean distance between s1[0] and s2[0], $$b$$ between s1[1] and s2[1] and $$c$$ between s1[2] and s2[2].

Attributes
feature

Feature object used to extract features from sequential data

is_order_invariant

Is this metric invariant to the sequence’s ordering

Methods

 are_compatible Checks if features can be used by metric.dist based on their shape. dist Computes a distance between two data points based on their features.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

VectorOfEndpointsFeature

class dipy.segment.metric.VectorOfEndpointsFeature

Bases: dipy.segment.featurespeed.CythonFeature

Extracts features from a sequential datum.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

The feature being extracted consists of one vector in the N-dimensional space pointing from one end-point of the sequence to the other (i.e. S[-1]-S[0]).

Attributes
is_order_invariant

Is this feature invariant to the sequence’s ordering

Methods

 extract Extracts features from a sequential datum. infer_shape Infers the shape of features extracted from a sequential datum.
__init__(*args, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

dist

dipy.segment.metric.dist()

Computes a distance between datum1 and datum2.

A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

Parameters
metricMetric object

Tells how to compute the distance between datum1 and datum2.

datum12D array

Sequence of N-dimensional points.

datum22D array

Sequence of N-dimensional points.

Returns
double

Distance between two data points.

distance_matrix

dipy.segment.metric.distance_matrix()

Computes the distance matrix between two lists of sequential data.

The distance matrix is obtained by computing the pairwise distance of all tuples spawn by the Cartesian product of data1 with data2. If data2 is not provided, the Cartesian product of data1 with itself is used instead. A sequence of N-dimensional points is represented as a 2D array with shape (nb_points, nb_dimensions).

Parameters
metricMetric object

Tells how to compute the distance between two sequential data.

data1list of 2D arrays

List of sequences of N-dimensional points.

data2list of 2D arrays

Llist of sequences of N-dimensional points.

Returns
2D array (double)

Distance matrix.

mdf

dipy.segment.metric.mdf(s1, s2)

Computes the MDF (Minimum average Direct-Flip) distance [Garyfallidis12] between two streamlines.

Streamlines must have the same number of points.

Parameters
s12D array

A streamline (sequence of N-dimensional points).

s22D array

A streamline (sequence of N-dimensional points).

Returns
double

Distance between two streamlines.

References

Garyfallidis12(1,2)

Garyfallidis E. et al., QuickBundles a method for tractography simplification, Frontiers in Neuroscience, vol 6, no 175, 2012.

otsu

dipy.segment.threshold.otsu(image, nbins=256)

Return threshold value based on Otsu’s method. Copied from scikit-image to remove dependency.

Parameters
imagearray

Input image.

nbinsint

Number of bins used to calculate histogram. This value is ignored for integer arrays.

Returns
thresholdfloat

Threshold value.

upper_bound_by_percent

dipy.segment.threshold.upper_bound_by_percent(data, percent=1)

Find the upper bound for visualization of medical images

Calculate the histogram of the image and go right to left until you find the bound that contains more than a percentage of the image.

Parameters
datandarray
percentfloat
Returns
upper_boundfloat

upper_bound_by_rate

dipy.segment.threshold.upper_bound_by_rate(data, rate=0.05)

Adjusts upper intensity boundary using rates

It calculates the image intensity histogram, and based on the rate value it decide what is the upperbound value for intensity normalization, usually lower bound is 0. The rate is the ratio between the amount of pixels in every bins and the bins with highest pixel amount

Parameters
datafloat

Input intensity value data

ratefloat

representing the threshold whether a spicific histogram bin that should be count in the normalization range

Returns
highfloat

the upper_bound value for normalization

ConstantObservationModel

class dipy.segment.tissue.ConstantObservationModel

Bases: object

Observation model assuming that the intensity of each class is constant. The model parameters are the means $$\mu_{k}$$ and variances $$\sigma_{k}$$ associated with each tissue class. According to this model, the observed intensity at voxel $$x$$ is given by $$I(x) = \mu_{k} + \eta_{k}$$ where $$k$$ is the tissue class of voxel $$x$$, and $$\eta_{k}$$ is a Gaussian random variable with zero mean and variance $$\sigma_{k}^{2}$$. The observation model is responsible for computing the negative log-likelihood of observing any given intensity $$z$$ at each voxel $$x$$ assuming the voxel belongs to each class $$k$$. It also provides a default parameter initialization.

Methods

 initialize_param_uniform Initializes the means and variances uniformly negloglikelihood Computes the gaussian negative log-likelihood of each class at each voxel of image assuming a gaussian distribution with means and variances given by mu and sigmasq, respectively (constant models along the full volume). prob_image Conditional probability of the label given the image seg_stats Mean and standard variation for N desired tissue classes update_param Updates the means and the variances in each iteration for all the labels.
__init__()

Initializes an instance of the ConstantObservationModel class

initialize_param_uniform

Initializes the means and variances uniformly

The means are initialized uniformly along the dynamic range of image. The variances are set to 1 for all classes

Parameters
imagearray

3D structural image

nclassesint

number of desired classes

Returns
muarray

1 x nclasses, mean for each class

sigmaarray

1 x nclasses, standard deviation for each class. Set up to 1.0 for all classes.

negloglikelihood

Computes the gaussian negative log-likelihood of each class at each voxel of image assuming a gaussian distribution with means and variances given by mu and sigmasq, respectively (constant models along the full volume). The negative log-likelihood will be written in nloglike.

Parameters
imagendarray

3D gray scale structural image

mundarray

mean of each class

sigmasqndarray

variance of each class

nclassesint

number of classes

Returns
nloglikendarray

4D negloglikelihood for each class in each volume

prob_image

Conditional probability of the label given the image

Parameters
imgndarray

3D structural gray-scale image

nclassesint

number of tissue classes

mundarray

1 x nclasses, current estimate of the mean of each tissue class

sigmasqndarray

1 x nclasses, current estimate of the variance of each tissue class

P_L_Nndarray

4D probability map of the label given the neighborhood.

Previously computed by function prob_neighborhood
Returns
P_L_Yndarray

4D probability of the label given the input image

seg_stats

Mean and standard variation for N desired tissue classes

Parameters
input_imagendarray

3D structural image

seg_imagendarray

3D segmented image

nclassint

number of classes (3 in most cases)

Returns
mu, varndarrays

1 x nclasses dimension Mean and variance for each class

update_param

Updates the means and the variances in each iteration for all the labels. This is for equations 25 and 26 of Zhang et. al., IEEE Trans. Med. Imag, Vol. 20, No. 1, Jan 2001.

Parameters
imagendarray

3D structural gray-scale image

P_L_Yndarray

4D probability map of the label given the input image computed by the expectation maximization (EM) algorithm

mundarray

1 x nclasses, current estimate of the mean of each tissue class.

nclassesint

number of tissue classes

Returns
mu_updndarray

1 x nclasses, updated mean of each tissue class

var_updndarray

1 x nclasses, updated variance of each tissue class

IteratedConditionalModes

class dipy.segment.tissue.IteratedConditionalModes

Bases: object

Methods

 icm_ising Executes one iteration of the ICM algorithm for MRF MAP estimation. initialize_maximum_likelihood Initializes the segmentation of an image with given prob_neighborhood Conditional probability of the label given the neighborhood Equation 2.18 of the Stan Z.
__init__()
icm_ising

Executes one iteration of the ICM algorithm for MRF MAP estimation. The prior distribution of the MRF is a Gibbs distribution with the Potts/Ising model with parameter beta:

https://en.wikipedia.org/wiki/Potts_model

Parameters
nloglikendarray

4D shape, nloglike[x, y, z, k] is the negative log likelihood of class k at voxel (x, y, z)

betafloat

positive scalar, it is the parameter of the Potts/Ising model. Determines the smoothness of the output segmentation.

segndarray

3D initial segmentation. This segmentation will change by one iteration of the ICM algorithm

Returns
new_segndarray

3D final segmentation

energyndarray

3D final energy

initialize_maximum_likelihood
Initializes the segmentation of an image with given

neg-loglikelihood

Initializes the segmentation of an image with neglog-likelihood field given by nloglike. The class of each voxel is selected as the one with the minimum neglog-likelihood (i.e. maximum-likelihood segmentation).

Parameters
nloglikendarray

4D shape, nloglike[x, y, z, k] is the likelihhood of class k for voxel (x, y, z)

Returns
segndarray

3D initial segmentation

prob_neighborhood

Conditional probability of the label given the neighborhood Equation 2.18 of the Stan Z. Li book (Stan Z. Li, Markov Random Field Modeling in Image Analysis, 3rd ed., Advances in Pattern Recognition Series, Springer Verlag 2009.)

Parameters
segndarray

3D tissue segmentation derived from the ICM model

betafloat

scalar that determines the importance of the neighborhood and the spatial smoothness of the segmentation. Usually between 0 to 0.5

nclassesint

number of tissue classes

Returns
PLNndarray

4D probability map of the label given the neighborhood of the voxel.

TissueClassifierHMRF

class dipy.segment.tissue.TissueClassifierHMRF(save_history=False, verbose=True)

Bases: object

This class contains the methods for tissue classification using the Markov Random Fields modeling approach

Methods

 classify(image, nclasses, beta[, tolerance, …]) This method uses the Maximum a posteriori - Markov Random Field approach for segmentation by using the Iterative Conditional Modes and Expectation Maximization to estimate the parameters.
__init__(save_history=False, verbose=True)

Initialize self. See help(type(self)) for accurate signature.

classify(image, nclasses, beta, tolerance=None, max_iter=None)

This method uses the Maximum a posteriori - Markov Random Field approach for segmentation by using the Iterative Conditional Modes and Expectation Maximization to estimate the parameters.

Parameters
imagendarray,

3D structural image.

nclassesint,

number of desired classes.

betafloat,

smoothing parameter, the higher this number the smoother the output will be.

tolerance: float,

value that defines the percentage of change tolerated to prevent the ICM loop to stop. Default is 1e-05.

max_iterfloat,

fixed number of desired iterations. Default is 100. If the user only specifies this parameter, the tolerance value will not be considered. If none of these two parameters

Returns
initial_segmentationndarray,

3D segmented image with all tissue types specified in nclasses.

final_segmentationndarray,

3D final refined segmentation containing all tissue types.

PVEndarray,

3D probability map of each tissue type.

dipy.segment.tissue.add_noise(signal, snr, S0, noise_type='rician')

Add noise of specified distribution to the signal from a single voxel.

Parameters
signal1-d ndarray

The signal in the voxel.

snrfloat

The desired signal-to-noise ratio. (See notes below.) If snr is None, return the signal as-is.

S0float

Reference signal for specifying snr.

noise_typestring, optional

The distribution of noise added. Can be either ‘gaussian’ for Gaussian distributed noise, ‘rician’ for Rice-distributed noise (default) or ‘rayleigh’ for a Rayleigh distribution.

Returns
signalarray, same shape as the input

Notes

SNR is defined here, following [1], as S0 / sigma, where sigma is the standard deviation of the two Gaussian distributions forming the real and imaginary components of the Rician noise distribution (see [2]).

References

1

Descoteaux, Angelino, Fitzgibbons and Deriche (2007) Regularized, fast and robust q-ball imaging. MRM, 58: 497-510

2

Gudbjartson and Patz (2008). The Rician distribution of noisy MRI data. MRM 34: 910-914.

Examples

>>> signal = np.arange(800).reshape(2, 2, 2, 100)
>>> signal_w_noise = add_noise(signal, 10., 100., noise_type='rician')