RandomAnisotropy

Simulation of an image with highly anisotropic spacing

`RandomAnisotropy`

Bases: RandomTransform

Downsample an image along an axis and upsample to initial space.

This transform simulates an image that has been acquired using anisotropic spacing and resampled back to its original spacing.

Similar to the work by Billot et al.: Partial Volume Segmentation of Brain MRI Scans of any Resolution and Contrast .

Parameters:

Name	Type	Description	Default
`axes`	`int \| tuple[int, ...]`	Axis or tuple of axes along which the image will be downsampled.	`(0, 1, 2)`
`downsampling`	`TypeRangeFloat`	Downsampling factor \(m \gt 1\). If a tuple \((a, b)\) is provided then \(m \sim \mathcal{U}(a, b)\).	`(1.5, 5)`
`image_interpolation`	`str`	Image interpolation used to upsample the image back to its initial spacing. Downsampling is performed using nearest neighbor interpolation. See Interpolation for supported interpolation types.	`'linear'`
`scalars_only`	`bool`	Apply only to instances of `torchio.ScalarImage`. This is useful when the segmentation quality needs to be kept, as in Billot et al. .	`True`
`**kwargs`		See `Transform` for additional keyword arguments.	`{}`

Examples:

>>> import torchio as tio
>>> transform = tio.RandomAnisotropy(axes=1, downsampling=2)
>>> transform = tio.RandomAnisotropy(
...     axes=(0, 1, 2),
...     downsampling=(2, 5),
... )   # Multiply spacing of one of the 3 axes by a factor randomly chosen in [2, 5]
>>> colin = tio.datasets.Colin27()
>>> transformed = transform(colin)

`call(data)`

Transform data and return a result of the same type.

Parameters:

Name	Type	Description	Default
`data`	`InputType`	Instance of `torchio.Subject`, 4D `torch.Tensor` or `numpy.ndarray` with dimensions \((C, W, H, D)\), where \(C\) is the number of channels and \(W, H, D\) are the spatial dimensions. If the input is a tensor, the affine matrix will be set to identity. Other valid input types are a SimpleITK image, a `torchio.Image`, a NiBabel Nifti1 image or a `dict`. The output type is the same as the input type.	required

`get_base_args()`

Provides easy access to the arguments used to instantiate the base class (Transform) of any transform.

This method is particularly useful when a new transform can be represented as a variant of an existing transform (e.g. all random transforms), allowing for seamless instantiation of the existing transform with the same arguments as the new transform during apply_transform.

Note

The p argument (probability of applying the transform) is excluded to avoid multiplying the probability of both existing and new transform.

`add_base_args(arguments, overwrite_on_existing=False)`

Add the init args to existing arguments

`validate_keys_sequence(keys, name)` `staticmethod`

Ensure that the input is not a string but a sequence of strings.

`to_hydra_config()`

Return a dictionary representation of the transform for Hydra instantiation.

RandomAnisotropy