eigenstrapping.geometry.geodesic_distmat

eigenstrapping.geometry.geodesic_distmat(surface, mask_input=None, use_cholmod=False, n_jobs=1, m=1.0)[source]
Compute geodesic distance using the heat diffusion method built into LaPy

Based on: doi/10.1145/2516971.2516977 Crane et al., “Geodesics in heat: A new approach to computing distance based on heat flow”

Parameters:

  • surf (str) – Path to input surface

  • mask_input (array_like or str) – Mask to remove from surf (i.e. medial wall) during the geodesic distance calculations

  • use_cholmod (bool, optional) – Specify whether to use scikit-sparse cholmod libraries (much faster), if False then uses scipy.sparse libraries instead (much slower). Default False.

  • n_jobs (int, optional) – Number of workers to use for parallel calls to ._thread_method(), default is 1.

  • m (float, optional) – Scaling factor for time in the heat kernel calculation. As m decreases, distances increase in accuracy (if you care) but increases computation time. Best put between 1.0 and 0.1. Default 1.0.

Returns:

D – Distance matrix of every vertex to every other vertex

Return type:

(N,N) np.ndarray