liom_toolkit.segmentation.volume_segmentation module

liom_toolkit.segmentation.volume_segmentation.fill_holes_2d_3d(mask: ndarray) → ndarray

Fill holes in a 2D and 3D mask. Source: https://github.com/linum-uqam/sbh-reconstruction/blob/51271c84347afccb21483cfd3fcbde77d537929c/slicercode/segmentation/brainMask.py

Parameters:

mask (np.ndarray) – The mask to fill holes in.

Returns:

The mask with holes filled.

Return type:

np.ndarray

liom_toolkit.segmentation.volume_segmentation.segment_3d_brain(volume: ndarray, k: int = 5, use_log: bool = True, threshold_method: str = 'otsu') → ndarray

Segment a 3D brain volume using a watershed algorithm. Source: https://github.com/linum-uqam/sbh-reconstruction/blob/51271c84347afccb21483cfd3fcbde77d537929c/slicercode/segmentation/brainMask.py

Parameters:

• volume (ants.ANTsImage) – The volume to segment.

• k (int) – The size of the median filter.

• use_log (bool) – Whether to use the log of the volume.

• threshold_method (str) – The threshold method to use. Either “otsu” or “triangle”.

Returns:

The segmented mask

Return type:

np.ndarray