Detect clusters with optional splitting

Usage

detect_clusters(
  data,
  threshold,
  method = c("watershed", "extract"),
  smooth = NULL,
  split_clusters = FALSE,
  peak_height = seq(0.05, 0.5, by = 0.05),
  min_new_cluster_size = 0.1,
  freqweight = 1,
  timeweight = 1,
  freq_axis = NULL,
  time_axis = NULL,
  data2 = NULL,
  dataRef = NULL,
  qval = NULL,
  filter_by_q = FALSE
)

Arguments

data: matrix of primary values.
threshold: numeric; binarization threshold.
method: "extract" or "watershed".
smooth: optional numeric or length-2 vector for Gaussian smoothing (pixels).
split_clusters: logical; attempt to split multi-peak clusters.
peak_height: numeric vector of relative peak heights for splitting.
min_new_cluster_size: minimum fraction of original cluster to keep split.
freqweight, timeweight: weights for distance reassignment.
freq_axis, time_axis: vectors for frequencies/times corresponding to rows/cols.
data2: optional secondary matrix for peak value extraction.
dataRef: optional reference matrix for computing reference statistics per cluster.
qval: optional FDR q for p-value adjustments.
filter_by_q: logical; if TRUE and qval is provided, keep only clusters with at least one pixel passing pAdj <= qval.

Value

list of cluster structs with coordinates and stats.