AFQ.recognition.clustering#

Functions#

`_compute_mean_euclidean_matrix`(group_n, group_m)
`_distance_to_similarity`(distance, sigmasq)
`_rectangular_similarity_matrix`(fgarray_sub, ...)
`spectral_atlas_label`(sub_fgarray, atlas_fgarray[, ...])	Use an existing atlas to label a new streamlines.
`subcluster_by_atlas`(sub_trk, mapping, dwi_ref, ...[, ...])	Use an existing atlas to label a new set of streamlines, and return the

Module Contents#

AFQ.recognition.clustering._compute_mean_euclidean_matrix(group_n, group_m)[source]#

AFQ.recognition.clustering._distance_to_similarity(distance, sigmasq)[source]#

AFQ.recognition.clustering._rectangular_similarity_matrix(fgarray_sub, fgarray_atlas, sigma)[source]#

AFQ.recognition.clustering.spectral_atlas_label(sub_fgarray, atlas_fgarray, atlas_data=None, sigma_multiplier=1.0, cluster_indices=None)[source]#

Use an existing atlas to label a new streamlines.

Parameters:

sub_fgarrayndarray: Resampled fiber group to be labeled.
atlas_fgarrayndarray: Resampled atlas to use for labelling.
atlas_datadict, optional: Precomputed atlas data formatted as a dictionary of arrays and floats. See afd.read_org800_templates as a reference.
sigma_multiplierfloat, optional: Multiplier for the sigma value used in computing the similarity matrix. Default is 1.0.
cluster_indiceslist of int, optional: If provided, only these cluster indices from the atlas will be used for labeling. Default is None, which uses all clusters.

Returns:

tuple of (ndarray, ndarray): Cluster indices for all the fibers and their embedding

AFQ.recognition.clustering.subcluster_by_atlas(sub_trk, mapping, dwi_ref, cluster_indices, atlas_data=None, n_points=20, batch_size=int(50000.0))[source]#

Use an existing atlas to label a new set of streamlines, and return the cluster indices for each streamline.

Parameters:

sub_trkStatefulTractogram: streamlines to be labeled.
mappingDIPY or pyAFQ mapping: Mapping to use to move streamlines.
dwi_refNifti1Image: Image defining reference for where the atlas streamlines move to.
cluster_indiceslist of int: Cluster indices from the atlas to use for labeling.
atlas_datadict, optional: Precomputed atlas data formatted as a dictionary of arrays and floats. See afd.read_org800_templates as a reference.
n_pointsint, optional: Number of points to resample streamlines to for labeling. Default is 20.
batch_sizeint, optional: Number of streamlines to process in a batch. Default is 50,000.