Cellpose cell segmentation — sg_segment

Segments cells using the Cellpose deep learning model via reticulate. Requires a working Python installation with cellpose installed. Use sg_setup_python() to configure the environment.

Usage

sg_segment_cellpose(
  image,
  model = c("cyto3", "cyto2", "nuclei", "tissuenet", "livecell", "custom"),
  channels = list(cytoplasm = 0L, nucleus = 1L),
  diameter = NULL,
  flow_threshold = 0.4,
  cellprob_threshold = 0,
  custom_model_path = NULL,
  batch_size = 8L,
  tile = TRUE
)

Arguments

image: An sg_image object.
model: Character string specifying the Cellpose model. One of "cyto3" (default), "cyto2", "nuclei", "tissuenet", "livecell", or "custom".
channels: A named list with elements cytoplasm and nucleus, each an integer channel index (0-based, following Cellpose convention). Default is list(cytoplasm = 0L, nucleus = 1L).
diameter: Numeric or NULL. Estimated cell diameter in pixels. If NULL, Cellpose estimates the diameter automatically.
flow_threshold: Numeric; flow error threshold for mask filtering. Default is 0.4.
cellprob_threshold: Numeric; cell probability threshold. Default is 0.0.
custom_model_path: Character string or NULL. Path to a custom-trained Cellpose model. Required when model = "custom".
batch_size: Integer; number of images to process in parallel on the GPU. Default is 8L.
tile: Logical; if TRUE (default), large images are processed in tiles to reduce memory usage.

Value

An sg_mask object with labelled cell regions.

Examples

# \donttest{
pixels <- array(runif(300), dim = c(10, 10, 3))
img <- new_sg_image(pixels, channels = c("R", "G", "B"))
# Requires Python + cellpose:
# mask <- sg_segment_cellpose(img, model = "cyto3")
# }