Labels

Fill in a module description here

compute_mfi

 compute_mfi (cell_mask, channel1, channel2=None, channel3=None,
              nuclear_mask=None)

Fast computation of mean fluorescence intensity (MFI) per cell and per nucleus using vectorized numpy.bincount operations.

	Type	Default	Details
cell_mask	2D int array		Label mask of cells (0 = background).
channel1	2D float array		First fluorescence channel.
channel2	NoneType	None	Second fluorescence channel.
channel3	NoneType	None	Third fluorescence channel.
nuclear_mask	NoneType	None	Label mask of nuclei (0 = background). If provided, nuclei MFI is computed.
Returns	pd.DataFrame		DataFrame with columns: - region: ‘cell’ or ‘nucleus’ - label: integer label - mfi_ch1, mfi_ch2, mfi_ch3 (if provided)

/opt/hostedtoolcache/Python/3.10.18/x64/lib/python3.10/site-packages/fastcore/docscrape.py:230: UserWarning: Unknown section Parameters:
  else: warn(msg)

source

plot_channel_mask

 plot_channel_mask (mask, channel1, channel2=None, channel3=None,
                    save_path=None, nuclear_mask=None)

*Plot up to three fluorescence channels with cell and nuclear mask overlays.

If only channel1 is provided, uses Spectral colormap. If channel2 and channel3 are provided, displays: - channel1 in green - channel2 in magenta - channel3 in yellow*

source

plot_label_mask

 plot_label_mask (mask, cell_type_map=None, type_colors=None,
                  default_color=(0, 0, 0), outline_color='black',
                  seed=None, save_path=None, nuclear_mask=None)

Plot a label mask coloring cells based on their assigned type. Unspecified or ‘other’ cells remain in default_color (black by default). Adds outlines for both cell and nuclear masks in specified color.

source

expand_with_cap

 expand_with_cap (label_image:numpy.ndarray, spacing:Tuple[float,float],
                  fixed_expand:float=10.0, max_area_ratio:float=1.5)

Expand labeled nuclei by up to a fixed radius, capped by max area ratio.

	Type	Default	Details
label_image	ndarray		2D integer array of labels (0=background).
spacing	Tuple		Physical size (µm) of each pixel along (row, column) axes.
fixed_expand	float	10.0	Nominal expansion radius in µm. Default is 10.0.
max_area_ratio	float	1.5	Maximum allowed area increase factor. Default is 1.5.
Returns	ndarray		Dilated label image, where each pixel is assigned to the nearest original label if within its per-pixel cap.

source

get_physical_size

 get_physical_size (filename:str, verbose:bool=False)

Compute the physical pixel size (µm/px) from a TIFF file’s resolution metadata.

	Type	Default	Details
filename	str		Path to the TIFF file.
verbose	bool	False	If True, print resolution details. Default is False.
Returns	float		Physical size of a pixel in microns. Raises if pixels are non-square.

source

plot_outlines

 plot_outlines (mask:numpy.ndarray, image:numpy.ndarray,
                save_path:str=None)

Overlay nuclear outlines from a mask onto an image and optionally save the result.

	Type	Default	Details
mask	ndarray		2D array of integer labels where nonzero values indicate nuclei.
image	ndarray		Grayscale image array of the same shape as `mask`.
save_path	str	None	File path in which to save the outline image. If None, no file is written.
Returns	matplotlib.figure.Figure		Figure object containing the outline overlay.

source

predict

 predict (img_list:list, use_gpu:bool, model_path:str=None,
          diameters:float=None, cellpose_param:tuple=(0.4, 0.0))

*Run a Cellpose segmentation model to generate segmentation masks for a list of images.

This function loads a Cellpose model and applies it to each image in the provided list. If a custom model path is given, it uses that pretrained model; otherwise, it defaults to the nuclei segmentation model. The segmentation thresholds can be controlled via cellpose_param, and an estimated diameter of the objects (nuclei) to segment can be provided when using the default model.*

	Type	Default	Details
img_list	list		A list of image arrays (grayscale) to segment.
use_gpu	bool		Indicates whether to use GPU acceleration for model evaluation.
model_path	str	None	Path to a custom pretrained Cellpose model. If None, the default nuclei model is used.
diameters	float	None	Estimated diameter (in pixels) of the objects (nuclei) to segment. This is used only when model_path is None.
cellpose_param	tuple	(0.4, 0.0)	A tuple (flow_threshold, cellprob_threshold) to control the Cellpose evaluation parameters. Defaults to (0.4, 0.0).
Returns	ndarray		Array of shape (N, H, W) containing the segmentation masks for each input image.