API Reference — Image

`swclib.image.swc2mask.Swc2Mask`

Converts an SWC neuron to a 3D binary mask volume.

Constructor

Swc2Mask(shape=(300, 300, 300), scale=(1, 1, 1), radius=1, method='line')

Parameter	Type	Default	Description
`shape`	`tuple[int, int, int]`	`(300,300,300)`	Output volume shape `(Z, Y, X)`
`scale`	`tuple[float, float, float]`	`(1,1,1)`	Voxel size `(sx, sy, sz)`
`radius`	`float`	`1`	Neurite rendering radius
`method`	`str`	`'line'`	Rendering method: `'line'` or `'sphere_cone'`

Methods

`run(swc, out_file=None) → ndarray`

Parameter	Type	Description
`swc`	`str` or `Swc`	SWC file path or object
`out_file`	`str` or `None`	If given, saves result as TIFF

Returns: ndarray of shape (Z, Y, X), dtype uint8.

Module-level functions

`swc_to_mask_line(swc, shape, scale, max_radius) → ndarray`

Distance-transform–based line rendering.

Parameter	Type	Description
`swc`	`Swc`	Source SWC object
`shape`	`tuple`	Output shape `(Z, Y, X)`
`scale`	`tuple`	Voxel size
`max_radius`	`float`	Cutoff radius

`swc_to_mask_sphere_cone(swc, shape, foreground_value, r_scale) → ndarray`

Sphere + truncated cone rendering.

Parameter	Type	Description
`swc`	`Swc`	Source SWC object
`shape`	`tuple`	Output shape `(Z, Y, X)`
`foreground_value`	`int`	Voxel value for foreground
`r_scale`	`float`	Radius scale factor

`swclib.image.mask2swc.Mask2Swc`

Converts a 3D binary mask to an SWC reconstruction.

Constructor

Mask2Swc(
    voxel_size=(1.0, 1.0, 1.0),
    connectivity=26,
    thres_fiber_min_len=20.0,
    thres_branch_min_len=15.0,
    thres_segment_merge_angle=None,
    thres_segment_merge_dist=None,
    smooth_window_size=None,
    smooth_poly=None,
    node_sample_distance=2.0,
)

Parameter	Type	Default	Description
`voxel_size`	`tuple`	`(1,1,1)`	Physical voxel size `(sx, sy, sz)` in µm
`connectivity`	`int`	`26`	Neighbor connectivity for skeleton: `6` or `26`
`thres_fiber_min_len`	`float`	`20.0`	Min length (µm) to keep an isolated fiber
`thres_branch_min_len`	`float`	`15.0`	Min length (µm) to keep a side branch
`thres_segment_merge_angle`	`float` or `None`	`None`	Max angle (°) for fiber merging
`thres_segment_merge_dist`	`float` or `None`	`None`	Max endpoint distance for merging
`smooth_window_size`	`int` or `None`	`None`	Savitzky-Golay smoothing window
`smooth_poly`	`int` or `None`	`None`	Savitzky-Golay polynomial order
`node_sample_distance`	`float`	`2.0`	Output node spacing after resampling

Methods

`run(mask, swc_path, soma_path=None, radius=0.1, verbose=False) → str`

Runs the full conversion pipeline and writes the SWC file.

Parameter	Type	Description
`mask`	`ndarray`	3D binary mask, shape `(Z, Y, X)`
`swc_path`	`str`	Output SWC file path
`soma_path`	`str` or `None`	Optional soma annotation file
`radius`	`float`	Default node radius in output SWC
`verbose`	`bool`	Print pipeline statistics

Returns: Output SWC file path (str).

`build_graph_from_skeleton(skel, connectivity, voxel_size) → nx.Graph`

Builds a NetworkX graph from a skeletonized binary mask.

`remove_isolated_nodes(G) → (nx.Graph, list)`

Removes degree-0 nodes.

`merge_fibers(G, voxel_size, ...) → (nx.Graph, int)`

Merges nearby near-collinear fiber endpoints. Returns (graph, num_merged).

`remove_short_fibers(G, voxel_size, thres_min_len) → (nx.Graph, int)`

Removes fibers shorter than the threshold.

`remove_short_branchs(G, voxel_size, thres_min_len) → (nx.Graph, int)`

Prunes side branches shorter than the threshold.

`refine_fibers(G, smooth_window_size, smooth_poly, node_sample_distance, voxel_size) → nx.Graph`

Applies Savitzky-Golay smoothing and uniform resampling to all fibers.