slice_extent_perturbation

Note

This page is a reference documentation. It only explains the function signature, and not how to use it. Please refer to the Habitat Guide and Python API guide (v2.0) for usage.

slice_extent_perturbation(mask: ndarray, grow_slices: int = 0, shrink_slices: int = 0, rng: Generator | None = None, max_slices: int = 0) → ndarray[source]

Add or remove whole axial slices at the superior/inferior ROI ends.

Models z-axis delineation variability: observers often agree in-plane but differ on the first and last slice they call tumour (slice-extent or “end-slice” disagreement). Operates on the z (first) axis only.

Provide either fixed grow_slices / shrink_slices or a random max_slices with rng (each end independently draws a count in [-max_slices, +max_slices]; positive grows, negative shrinks).

Parameters:

  • mask – Integer / boolean label array in (z, y, x) order.

  • grow_slices – Number of slices to append at each occupied end by copying the nearest occupied slice’s labels.

  • shrink_slices – Number of occupied slices to remove at each end.

  • rng – Random generator enabling random per-end counts; requires max_slices > 0.

  • max_slices – Bound for the random per-end slice count.

Returns:

A new label array, same shape and dtype as mask.

Raises:

ValueError – If shrinking removes every occupied slice, or if both fixed and random modes are mixed inconsistently.