uniq_per_pix#

crispy.grid_ridge.uniq_per_pix(coord, mask, coord_in_xfirst=False, start_index=1)[source]#

Reduce a list of ridge coordinates to one unique point per pixel.

Processes ridge coordinates to retain a single representative point per pixel based on a provided binary mask. The representative point is selected as the one with the median value along the last coordinate axis.

Parameters:

  • coord (ndarray) – Ridge coordinates, shape (D, n), where D is the number of dimensions (e.g., 2 or 3) and n is the number of points.

  • mask (ndarray) – Binary mask array, shape matching the reference grid, where True indicates pixels of interest.

  • coord_in_xfirst (bool, optional, default=False) – If True, assumes the input coordinates are ordered with x as the first axis. If False, assumes z is the first axis for 3D data or y for 2D data.

  • start_index (int, optional, default=1) – Starting index for the coordinate system. Adjusts the input coordinates before processing.

Returns:

coord_uniq – Reduced set of coordinates, shape (D, m), where m is the number of unique pixels with a representative coordinate.

Return type:

ndarray

Notes

  • This function is optimized for cases where the input mask represents gridded, one-voxel-wide skeletons or spines.

  • The median value along the last axis (e.g., z in 3D) is used to select the representative point for each pixel.

Examples

Reduce ridge coordinates to one per pixel:

>>> import numpy as np
>>> from crispy import grid_ridge
>>> coords = np.array([[0, 1, 1, 2], [0, 0, 0, 0], [0, 0, 1, 1]])  # 3D coordinates
>>> mask = np.zeros((3, 3, 3), dtype=bool)
>>> mask[1, 1, 0] = True
>>> mask[1, 1, 1] = True
>>> reduced_coords = grid_ridge.uniq_per_pix(coords, mask)
>>> print(reduced_coords)
[[1]
 [1]
 [0]]
On this page

This Page