volume

voxel.volume.stack(*vols) -> Volume

Concatenate (stack) multiple volumes channel-wise. Assumes the volumes are in the same image space (with the same base shape).

Parameters:

• *vols (Volume, default: () ) –

  Volumes to merge.

Returns:

• Volume –

  Single channel-stacked volume instance.

voxel.volume.volume_grid(baseshape, transform=None, localshape=None, device=None) -> torch.Tensor

Construct a grid of 3D voxel coordinates of the shape (W, H, D, 3).

Parameters:

• baseshape (Size) –

  Spatial (3D) shape of the volume grid.

• transform (AffineMatrix, default: None ) –

  Grid voxel coordinate transform.

• localshape (Size, default: None ) –

  If provided, the grid is normalized to the range [-1, 1] using this spatial shape and the coordinate order is swapped (for torch sampling methods).

• device (device, default: None ) –

  Device on which to allocate the grid data.

Returns:

• Tensor –

  Grid volume tensor.

voxel.volumes_equal(a, b, vol_tol=1e-06, geom_tol=1e-06) -> bool

Check if two volumes are equal within a given tolerance.

Parameters:

• a (Volume) –

  First volume to compare.

• b (Volume) –

  Second volume to compare.

• vol_tol (float, default: 1e-06 ) –

  Absolute tolerance for volume tensor comparison.

• geom_tol (float, default: 1e-06 ) –

  Absolute tolerance for geometry comparison.

Returns:

• bool –

  True if the volumes are equal, False otherwise.