tasks.stardist_segmentation
tasks.stardist_segmentation
Functions
Name | Description |
---|---|
stardist_segmentation | Segment cells with Stardist |
stardist_segmentation
tasks.stardist_segmentation.stardist_segmentation(zarr_url, channel, roi_table='FOV_ROI_table', stardist_model='2D_versatile_fluo', label_name='nuclei', prob_thresh=None, nms_thresh=None, scale=1, level=0, overwrite=False)
Segment cells with Stardist
Parameters
Name | Type | Description | Default |
---|---|---|---|
zarr_url |
str | Path to an OME-ZARR Image | required |
channel |
ChannelInputModel | Channel for segmentation; requires either wavelength_id (e.g. A01_C01 ) or label (e.g. DAPI ) but not both |
required |
roi_table |
str | Name of the ROI table | 'FOV_ROI_table' |
stardist_model |
str | Name of the Stardist model (“2D_versatile_fluo”, “2D_versatile_he”, “2D_demo”, “3D_demo”) | '2D_versatile_fluo' |
label_name |
str | Name of the labels folder | 'nuclei' |
prob_thresh |
Optional[float] | prob_thresh: Only consider objects with predicted object probability above this threshold | None |
nms_thresh |
Optional[float] | Perform non-maximum suppression (NMS) that considers two objects to be the same when their area/surface overlap exceeds this threshold | None |
scale |
int | Scale the input image internally by a factor and rescale the output accordingly. | 1 |
level |
int | Resolution level (0 = full resolution) | 0 |
overwrite |
bool | Whether to overwrite any existing OME-ZARR segmentations | False |