platipy.imaging.generation package#

Submodules#

platipy.imaging.generation.augment module#

class platipy.imaging.generation.augment.ContractAugment(mask, vector_contract=(10, 10, 10), gaussian_smooth=5, bone_mask=False)#

Bases: DeformableAugment

augment()#
class platipy.imaging.generation.augment.DeformableAugment#

Bases: ABC

abstract augment()#
class platipy.imaging.generation.augment.ExpandAugment(mask, vector_expand=(10, 10, 10), gaussian_smooth=5, bone_mask=False)#

Bases: DeformableAugment

augment()#
class platipy.imaging.generation.augment.ShiftAugment(mask, vector_shift=(10, 10, 10), gaussian_smooth=5)#

Bases: DeformableAugment

augment()#
platipy.imaging.generation.augment.apply_augmentation(image, augmentation, masks=[])#
platipy.imaging.generation.augment.generate_random_augmentation(ct_image, masks)#

platipy.imaging.generation.dvf module#

platipy.imaging.generation.dvf.generate_field_asymmetric_contract(mask_image, vector_asymmetric_contract=(10, 10, 10), gaussian_smooth=5, compute_real_dvf=False)#

Contracts a structure (defined using a binary mask) using a specified vector.

Parameters:

  • mask_image ([SimpleITK.Image]) – The binary mask to contract.

  • vector_asymmetric_contract (tuple, optional) – The contraction vector applied to the entire binary mask. Convention: (+/-, +/-, +/-) = (sup/inf, post/ant, left/right) border is contracted. Defined in millimetres. Defaults to (10, 10, 10).

  • gaussian_smooth (int | list, optional) – Scale of a Gaussian kernel used to smooth the deformation vector field. Defaults to 5.

  • compute_real_dvf (bool, optional) – If True, the real deformation vector field is computed. This involves a slower computation. Defaults to False.

Returns:

The binary mask following the contract. [SimpleITK.DisplacementFieldTransform]: The transform representing the contract. [SimpleITK.Image]: The displacement vector field representing the contract.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.dvf.generate_field_asymmetric_extend(mask_image, vector_asymmetric_extend=(10, 10, 10), gaussian_smooth=5)#

Extends a structure (defined using a binary mask) using a specified vector.

Parameters:

  • mask_image ([SimpleITK.Image]) – The binary mask to extend.

  • vector_asymmetric_extend (tuple, optional) – The extension vector applied to the entire binary mask. Convention: (+/-, +/-, +/-) = (sup/inf, post/ant, left/right) border is extended. Defined in millimetres. Defaults to (10, 10, 10).

  • gaussian_smooth (int | list, optional) – Scale of a Gaussian kernel used to smooth the deformation vector field. Defaults to 5.

Returns:

The binary mask following the extension. [SimpleITK.DisplacementFieldTransform]: The transform representing the extension. [SimpleITK.Image]: The displacement vector field representing the extension.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.dvf.generate_field_expand(mask, bone_mask=False, expand=3, gaussian_smooth=5, use_internal_deformation=True)#

Expands a structure (defined using a binary mask) using a specified vector to define the dilation kernel.

Parameters:

  • mask ([SimpleITK.Image]) – The binary mask to expand.

  • bone_mask ([SimpleITK.Image, optional]) – A binary mask defining regions where we expect restricted deformations.

  • vector_asymmetric_extend (int |tuple, optional) –

    The expansion vector applied to the entire

    binary mask.

    Convention: (z,y,x) size of expansion kernel. Defined in millimetres. Defaults to 3.

  • gaussian_smooth (int | list, optional) – Scale of a Gaussian kernel used to smooth the deformation vector field. Defaults to 5.

Returns:

The binary mask following the expansion. [SimpleITK.DisplacementFieldTransform]: The transform representing the expansion. [SimpleITK.Image]: The displacement vector field representing the expansion.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.dvf.generate_field_radial_bend(reference_image, body_mask, reference_point, axis_of_rotation=[0, 0, -1], scale=0.1, mask_bend_from_reference_point=('z', 'inf'), gaussian_smooth=5)#

Generates a synthetic field characterised by radial bending. Typically, this field would be used to simulate a moving head and so masking is important.

Parameters:

  • reference_image ([SimpleITK.Image]) – The image to be deformed.

  • body_mask ([SimpleITK.Image]) – A binary mask in which the deformation field will be defined

  • reference_point ([tuple]) – The point (z,y,x) about which the rotation field is defined.

  • axis_of_rotation (tuple, optional) – The axis of rotation (z,y,x). Defaults to [0, 0, -1].

  • scale (int, optional) – The deformation vector length at each point will equal scale multiplied by the distance to that point from reference_point. Defaults to 1.

  • mask_bend_from_reference_point (tuple, optional) – The dimension (z=axial, y=coronal, x=sagittal) and limit (inf/sup, post/ant, left/right) for masking the vector field, relative to the reference point. Defaults to (“z”, “inf”).

  • gaussian_smooth (int | list, optional) – Scale of a Gaussian kernel used to smooth the deformation vector field. Defaults to 5.

Returns:

The binary mask following the expansion. [SimpleITK.DisplacementFieldTransform]: The transform representing the expansion. [SimpleITK.Image]: The displacement vector field representing the expansion.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.dvf.generate_field_shift(mask_image, vector_shift=(10, 10, 10), gaussian_smooth=5)#

Shifts (moves) a structure defined using a binary mask.

Parameters:

  • mask_image ([SimpleITK.Image]) – The binary mask to shift.

  • vector_shift (tuple, optional) – The displacement vector applied to the entire binary mask. Convention: (+/-, +/-, +/-) = (sup/inf, post/ant, left/right) shift. Defined in millimetres. Defaults to (10, 10, 10).

  • gaussian_smooth (int | list, optional) – Scale of a Gaussian kernel used to smooth the deformation vector field. Defaults to 5.

Returns:

The binary mask following the shift. [SimpleITK.DisplacementFieldTransform]: The transform representing the shift. [SimpleITK.Image]: The displacement vector field representing the shift.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.image module#

platipy.imaging.generation.image.insert_cylinder(arr, cyl_radius=4, cyl_height=2, cyl_centre=(0, 0, 0))#

Insert a cylinder into the given array. The cylinder vertical extent is +/- 0.5 * height

Parameters:

  • arr (np.ndarray) – The array into which the cylinder is inserted

  • cyl_radius (int, optional) – Cylinder radius. Defaults to 4.

  • cyl_height (int, optional) – Cylinder height. Defaults to 2.

  • cyl_centre (tuple, optional) – Cylinder centre. Defaults to (0, 0, 0).

Returns:

The original array with a cylinder (value 1)

Return type:

np.ndarray

platipy.imaging.generation.image.insert_cylinder_image(image, cyl_radius=(5, 5), cyl_height=10, cyl_centre=(0, 0, 0))#

Insert a cylinder into an image

Parameters:

  • image (SimpleITK.Image) –

  • cyl_radius (tuple, optional) – Cylinder radius, can be defined as a single value or a tuple (will generate an ellipsoid). Defaults to (5,5).

  • cyl_height (int, optional) – Cylinder height. Defaults to 10.

  • cyl_centre (tuple, optional) – Cylinder centre. Defaults to (0,0,0).

Returns:

Image with cylinder inserted

Return type:

SimpleITK.Image

platipy.imaging.generation.image.insert_sphere(arr, sp_radius=4, sp_centre=(0, 0, 0))#

Insert a sphere into the given array

Parameters:

  • arr (np.array) – Array in which to insert sphere

  • sp_radius (int, optional) – The radius of the sphere. Defaults to 4.

  • sp_centre (tuple, optional) – The position at which the sphere should be inserted. Defaults to (0, 0, 0).

Returns:

An array with the sphere inserted

Return type:

np.array

platipy.imaging.generation.image.insert_sphere_image(image, sp_radius, sp_centre)#

Insert a sphere into an image

Parameters:

  • image (sitk.Image) – Image in which to insert sphere

  • sp_radius (int | list, optional) – The radius of the sphere. Can also be defined as a vector. Defaults to 4.

  • sp_centre (tuple, optional) – The position at which the sphere should be inserted. Defaults to (0, 0, 0).

Returns:

An array with the sphere inserted

Return type:

np.array

platipy.imaging.generation.mask module#

platipy.imaging.generation.mask.extend_mask(mask, direction=('ax', 'sup'), extension_mm=10, interior_mm_shape=10)#

Extends a binary label (mask) a number of slices. PROTOTYPE! Currently can only extend in axial directions (superior of inferior). The shape of the extended part is based on some number of interior slices, as defined.

Parameters:

  • mask (SimpleITK.Image) – The input binary label (mask).

  • direction (tuple, optional) – The direction as a tuple. First element is axis, second element is direction. Defaults to (“ax”, “sup”).

  • extension_mm (int, optional) – The extension in millimeters. Defaults to 10.

  • interior_mm_shape (int, optional) – The length on which to base the extension shape. Defaults to 10.

Returns:

The output (extended mask).

Return type:

SimpleITK.Image

platipy.imaging.generation.mask.get_bone_mask(image, lower_threshold=350, upper_threshold=3500, max_hole_size=5)#

Automatically generate a binary mask of bones from a CT image.

Parameters:

  • image ([SimpleITK.Image]) – The patient x-ray CT image to segment.

  • lower_threshold (int, optional) – Lower voxel value for threshold. Defaults to 350.

  • upper_threshold (int, optional) – Upper voxel value for threshold. Defaults to 3500.

  • max_hole_size (int | list | bool, optional) – Maximum hole size to be filled in millimetres. Can be specified as a vector (z,y,x). Defaults to 5.

Returns:

The binary bone mask.

Return type:

[SimpleITK.Image]

platipy.imaging.generation.mask.get_external_mask(image, lower_threshold=-100, upper_threshold=2500, dilate=1, max_hole_size=False)#

Automatically generate a binary mask of the patient external contour. Uses slice-wise convex hull generation.

Parameters:

  • image ([SimpleITK.Image]) – The patient x-ray CT image to segment. May work with other modalities with modified thresholds.

  • lower_threshold (int, optional) – Lower voxel value for threshold. Defaults to -100.

  • upper_threshold (int, optional) – Upper voxel value for threshold. Defaults to 2500.

  • dilate (int | list | bool, optional) – Dilation filter size applied to the binary mask. Can be specified as a vector (z,y,x). Defaults to 1.

  • max_hole_size (int | list | bool, optional) – Maximum hole size to be filled in millimetres. Can be specified as a vector (z,y,x). Defaults to False.

Returns:

The binary external mask.

Return type:

[SimpleITK.Image]

Module contents#