napari_cellseg3d.code_models.instance_segmentation

Instance segmentation methods for 3D images.

Module Attributes

USE_SLIDING_WINDOW

If True, uses a sliding window to perform instance segmentation to avoid memory issues.

Functions

binary_connected(volume[, thres, thres_small])	Convert binary foreground probability maps to instance masks via connected-component labeling.
binary_watershed(volume[, thres_objects, ...])	Convert binary foreground probability maps to instance masks via watershed segmentation algorithm.
clear_large_objects(image[, ...])	Uses watershed to label all obejcts, and removes the ones with a volume larger than the specified threshold.
clear_small_objects(image, threshold[, ...])	Calls skimage.remove_small_objects to remove small fragments that might be artifacts.
threshold(volume, thresh)	Remove all values smaller than the specified threshold in the volume.
to_semantic(image[, is_file_path])	Converts a ground-truth label to semantic (binary 0/1) labels.
volume_stats(volume_image)	Computes various statistics from instance labels and returns them in a dict.
voronoi_otsu(volume, spot_sigma, outline_sigma)	Voronoi-Otsu labeling from pyclesperanto.

Classes

ConnectedComponents([widget_parent])	Widget class for Connected Components instance segmentation.
ImageStats(volume, centroid_x, centroid_y, ...)	Dataclass containing various statistics from instance labels.
InstanceMethod(name, function, num_sliders, ...)	Base class for instance segmentation methods.
InstanceWidgets([parent])	Base widget with several sliders, for use in instance segmentation parameters.
VoronoiOtsu([widget_parent])	Widget class for Voronoi-Otsu labeling from pyclesperanto.
Watershed([widget_parent])	Widget class for Watershed segmentation.