fastgac.IcoCharts

class fastgac.IcoCharts

Contains charts of an unwrapped Icosahedron. This is basically my implementation of unwrapping an icosahedron as described in: Gauge Equivariant Convolutional Networks and the Icosahedral CNN - https://arxiv.org/abs/1902.04615

__init__(self: fastgac_pybind.IcoCharts, level: int = 1) → None

fill_image(self: fastgac_pybind.IcoCharts, normalized_vertex_count: fastgac_pybind.VectorDouble) → None

Fills the the image using the normalized vertex counts

find_peaks(self: fastgac_pybind.IcoCharts, threshold_abs: int = 25, exclude_border: bool = False) → List[List[int[2]]]

Finds all peaks inside the image

property image

Returns an unwrapped image of the IcoCharts

property image_to_vertex_idx

Fast lookup matrix for image creation. Each pixel hold the icosahedron vertex index it corresponds to

property mask

Boolean mask corresponding to valid cells. False(0) corresponds to ghost/halo cells

property sphere_mesh

The full icosahedron the IcoChart is unwrapping