Deep learning for NeuroImaging in Python.

Note

This page is a reference documentation. It only explains the class signature, and not how to use it. Please refer to the gallery for the big picture.

class surfify.augmentation.base.SurfCutOut(vertices, triangles, neighs=None, patch_size=3, n_patches=1, sigma=0, replacement_value=0, cachedir=None)[source]¶

Starting from random vertices, the SurfCutOut sets an adaptive connex neighborhood to zero.

See also

surfify.utils.neighbors

Init class.

Parameters:

vertices : array (N, 3)

icosahedron’s vertices.

triangles : array (M, 3)

icosahdron’s triangles.

neighs : dict, default None

optionnaly specify the DiNe neighboors of each vertex as build with sufify.utils.neighbors, ie. a dictionary with vertices row index as keys and a dictionary of neighbors vertices row indexes organized by rings as values.

patch_size : int, default 3

the number of neighboring rings from one node to be considered during the ablation.

n_patches : int, default 1

the number of patches to be considered.

sigma : int, default 0

use different patch size in [patch_size-sigma, patch_size+sigma] for each cutout.

replacement_value : float, default 0

the replacement patch value.

cachedir : str, default None

the optional path to cache the neighbors function output.

run(data)[source]¶

Applies the cut out (ablation) augmentation to the data.

Parameters:: data : array (N, )

input data/texture.
Returns:: data : arr (N, )

ablated input data.