Build a contourlet-like directional support mask from a spectral cube

This function extends the default starlet support with a directional, multiscale arm-sensitive support map. The implementation is contourlet-like rather than a full contourlet transform: it uses a bank of oriented, elongated Gaussian filters across multiple scales, combines them into a directional response map, and optionally grows the starlet support into nearby arm-like pixels.

Usage

build_contourlet_mask(
  input,
  collapse_fn = collapse_white_light,
  pretransform = "none",
  combine = c("hybrid", "standalone"),
  base_mask = NULL,
  starlet_J = 5,
  starlet_scales = 2:5,
  include_coarse = FALSE,
  denoise_k = 2.5,
  mode = c("soft", "hard"),
  positive_only = TRUE,
  angles_deg = seq(0, 165, by = 15),
  scales = list(c(4.5, 1.2), c(7.5, 1.8), c(11, 2.6)),
  background_sigma = 8,
  score_low_q = 0.92,
  score_high_q = 0.98,
  contrast_low_q = 0.8,
  contrast_high_q = 0.93,
  max_growth_iter = 8L,
  close_size = 3L,
  open_size = 3L
)

Arguments

input: 3-D array or FITS-like list with imDat.
collapse_fn: Function used to collapse the cube to a 2-D image.
pretransform: Optional spectral pretransform applied to the cube before the white-light collapse.
combine: Either "hybrid" (recommended; start from the starlet support and extend it directionally) or "standalone" (use only the directional support).
base_mask: Optional precomputed base mask. When combine = "hybrid" and base_mask is NULL, the base support is built from build_starlet_mask().
starlet_J: Number of starlet scales used when building the hybrid base support.
starlet_scales: Scales to reconstruct for the hybrid base support.
include_coarse: Logical; include the coarse plane in the hybrid base support reconstruction.
denoise_k: Optional starlet denoising threshold used in the hybrid base support.
mode: Thresholding mode for the hybrid base support reconstruction.
positive_only: Logical; keep only positive reconstruction values in the hybrid base support.
angles_deg: Vector of filter orientations in degrees.
scales: List of (sigma_major, sigma_minor) pairs defining the directional filter bank.
background_sigma: Smoothing scale used to estimate the local background before the directional response is measured.
score_low_q: Lower quantile used to define the allowed directional growth region.
score_high_q: Upper quantile used to define strong directional seeds.
contrast_low_q: Lower quantile used to define the allowed contrast growth region.
contrast_high_q: Upper quantile used to define strong contrast seeds.
max_growth_iter: Maximum number of constrained growth iterations.
close_size: Size of the binary closing brush.
open_size: Size of the binary opening brush.

Value

A list containing the collapsed image, the optional base support, the final logical mask, the directional score map, the local contrast map, and threshold metadata.