Add support to invert RescaleIntensity transform

src/torchio/transforms/preprocessing/intensity/rescale.py

@@ -1,6 +1,5 @@
 import warnings
 from typing import Optional
-from typing import Tuple
 
 import numpy as np
 import torch
@@ -48,7 +47,7 @@ def __init__(
  out_min_max: TypeRangeFloat = (0, 1),
  percentiles: TypeRangeFloat = (0, 100),
  masking_method: TypeMaskingMethod = None,
- in_min_max: Optional[Tuple[float, float]] = None,
+ in_min_max: Optional[TypeRangeFloat] = None,
  **kwargs
  ):
  super().__init__(masking_method=masking_method, **kwargs)
@@ -60,7 +59,21 @@ def __init__(
  self.percentiles = self._parse_range(
  percentiles, 'percentiles', min_constraint=0, max_constraint=100,
  )
- self.args_names = ['out_min_max', 'percentiles', 'masking_method']
+
+ self.in_min: Optional[float]
+ self.in_max: Optional[float]
+ if self.in_min_max is not None:
+ self.in_min, self.in_max = self._parse_range(
+ self.in_min_max, 'in_min_max',
+ )
+ else:
+ self.in_min = None
+ self.in_max = None
+
+ self.args_names = [
+ 'out_min_max', 'percentiles', 'masking_method', 'in_min_max',
+ ]
+ self.invert_transform = False
 
  def apply_normalization(
  self,
@@ -91,20 +104,29 @@ def rescale(
  cutoff = np.percentile(values, self.percentiles)
  np.clip(array, *cutoff, out=array) # type: ignore[call-overload]
  if self.in_min_max is None:
- in_min, in_max = array.min(), array.max()
- else:
- in_min, in_max = self.in_min_max
- in_range = in_max - in_min
+ self.in_min_max = self._parse_range(
+ (array.min(), array.max()), 'in_min_max',
+ )
+ self.in_min, self.in_max = self.in_min_max
+ assert self.in_min is not None
+ assert self.in_max is not None
+ in_range = self.in_max - self.in_min
  if in_range == 0: # should this be compared using a tolerance?
  message = (
  f'Rescaling image "{image_name}" not possible'
  ' because all the intensity values are the same'
  )
  warnings.warn(message, RuntimeWarning, stacklevel=2)
  return tensor
- array -= in_min
- array /= in_range
  out_range = self.out_max - self.out_min
- array *= out_range
- array += self.out_min
+ if self.invert_transform:
+ array -= self.out_min
+ array /= out_range
+ array *= in_range
+ array += self.in_min
+ else:
+ array -= self.in_min
+ array /= in_range
+ array *= out_range
+ array += self.out_min
  return torch.as_tensor(array)

tests/transforms/preprocessing/test_rescale.py

@@ -107,3 +107,14 @@ def test_empty_mask(self):
  rescale = tio.RescaleIntensity(masking_method='label')
  with pytest.warns(RuntimeWarning):
  rescale(subject)
+
+ def test_invert_rescaling(self):
+ torch.manual_seed(0)
+ transform = tio.RescaleIntensity(out_min_max=(0, 1))
+ data = torch.rand(1, 2, 3, 4).double()
+ subject = tio.Subject(t1=tio.ScalarImage(tensor=data))
+ transformed = transform(subject)
+ assert transformed.t1.data.min() == 0
+ assert transformed.t1.data.max() == 1
+ inverted = transformed.apply_inverse_transform()
+ self.assert_tensor_almost_equal(inverted.t1.data, data)