napari_prism.im.segment_tma

Contents

napari_prism.im.segment_tma#

napari_prism.im.segment_tma(spatialdata, image_name, output_segmentation_label, segmentation_channel, tiling_shapes=None, model_type='nuclei', nuclei_diam_um=None, channel_merge_method='max', optional_nuclear_channel=None, tiling_shapes_annotation_column=None, normalize=True, cellprob_threshold=0.0, flow_threshold=0.4, custom_model=False, denoise_model=None, preview=False, reference_coordinate_system='global', scale='scale0', inplace=True)#

Performs cell segmentation using Cellpose to the given image.

Parameters:

  • spatialdata (SpatialData) – The spatialdata object to segment.

  • image_name (str) – The name of the image to segment.

  • output_segmentation_label (str) – The name of the output segmentation label.

  • segmentation_channel (str | list[str]) – The channel or channels to segment.

  • tiling_shapes (Union[GeoDataFrame, str, None] (default: None)) – Shapes to explicitly tile by. Usually this will be the shapes denoting TMA core regions.

  • model_type (Literal['cyto3', 'cyto2', 'cyto', 'nuclei', 'tissuenet_cp3', 'livecell_cp3', 'yeast_PhC_cp3', 'yeast_BF_cp3', 'bact_phase_cp3', 'bact_fluor_cp3', 'deepbacs_cp3', 'cyto2_cp3'] (default: 'nuclei')) – The type of cellpose model to use.

  • nuclei_diam_um (Optional[float] (default: None)) – The diameter of the nuclei in microns. If this is negative, then the diameter is automatically estimated.

  • channel_merge_method (Literal['max', 'mean', 'sum', 'median'] (default: 'max')) – The method to merge multiple channels. Options: “max”, “mean”, “sum”, “median”.

  • optional_nuclear_channel (Optional[str] (default: None)) – The optional nuclear channel (if supported).

  • tiling_shapes_annotation_column (Optional[str] (default: None)) – The column in tiling_shapes to annotate distinct regions by. This is added as a label to each cell.

  • normalize (bool (default: True)) – If True, does image intensity normalization.

  • cellprob_threshold (float (default: 0.0)) – The cell probability threshold.

  • flow_threshold (float (default: 0.4)) – The flow threshold.

  • custom_model (Path | str | bool (default: False)) – The custom model to use, if supplied.

  • denoise_model (Optional[Literal['nan', 'denoise_cyto3', 'deblur_cyto3', 'upsample_cyto3', 'oneclick_cyto3', 'denoise_cyto2', 'deblur_cyto2', 'upsample_cyto2', 'oneclick_cyto2', 'denoise_nuclei', 'deblur_nuclei', 'upsample_nuclei', 'oneclick_nuclei']] (default: None)) – The denoiser model to use.

  • preview (bool (default: False)) – If True, only returns the input image (i.e. everything up to just before segmentation).

  • reference_coordinate_system (str (default: 'global')) – The reference coordinate system to use.

  • scale (default: 'scale0') – The scale to segment.

  • inplace (bool (default: True)) – If True, modifies the spatialdata object in place. If False, returns the segmentation mask and transformation sequence.

Return type:

SpatialData | DataArray | tuple[DataArray, BaseTransformation]

Returns:

If inplace is True, returns the modified spatialdata object. If False, returns a tuple of the segmentation mask and transformation sequence.

Contents