ENH: Adding --float to antsReg and BFit cmdlines

Expose --float option in ants registration.  This allows
for computations to be done in single precision rather
than double precision.

Author: hjmjohnson

nipype/interfaces/ants/registration.py

@@ -291,6 +291,11 @@ class RegistrationInputSpec(ANTSCommandInputSpec):
           'each stage directly updates the estimated linear transform from the previous '
           'stage. (e.g. Translation -> Rigid -> Affine). '
           ))
+    # NOTE: Even though only 0=False and 1=True are allowed, ants uses integer
+    # values instead of booleans
+    float = traits.Bool(
+        argstr='--float %d', default=False,
+        desc=('Use float instead of double for computations.'))
 
     transforms = traits.List(traits.Enum('Rigid', 'Affine', 'CompositeAffine',
                                          'Similarity', 'Translation', 'BSpline',
@@ -405,6 +410,16 @@ class Registration(ANTSCommand):
     >>> reg3.cmdline
     'antsRegistration --collapse-output-transforms 0 --dimensionality 3 --initial-moving-transform [ trans.mat, 1 ] --initialize-transforms-per-stage 0 --interpolation Linear --output [ output_, output_warped_image.nii.gz ] --transform Affine[ 2.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32, Random, 0.05 ] --convergence [ 1500x200, 1e-08, 20 ] --smoothing-sigmas 1.0x0.0vox --shrink-factors 2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --transform SyN[ 0.25, 3.0, 0.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32 ] --convergence [ 100x50x30, 1e-09, 20 ] --smoothing-sigmas 2.0x1.0x0.0vox --shrink-factors 3x2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --winsorize-image-intensities [ 0.025, 0.975 ]  --write-composite-transform 1'
 
+    >>> reg3a = copy.deepcopy(reg)
+    >>> reg3a.inputs.float = True
+    >>> reg3a.cmdline
+    'antsRegistration --collapse-output-transforms 0 --dimensionality 3 --float 1 --initial-moving-transform [ trans.mat, 1 ] --initialize-transforms-per-stage 0 --interpolation Linear --output [ output_, output_warped_image.nii.gz ] --transform Affine[ 2.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32, Random, 0.05 ] --convergence [ 1500x200, 1e-08, 20 ] --smoothing-sigmas 1.0x0.0vox --shrink-factors 2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --transform SyN[ 0.25, 3.0, 0.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32 ] --convergence [ 100x50x30, 1e-09, 20 ] --smoothing-sigmas 2.0x1.0x0.0vox --shrink-factors 3x2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --winsorize-image-intensities [ 0.0, 1.0 ]  --write-composite-transform 1'
+
+    >>> reg3b = copy.deepcopy(reg)
+    >>> reg3b.inputs.float = False
+    >>> reg3b.cmdline
+    'antsRegistration --collapse-output-transforms 0 --dimensionality 3 --float 0 --initial-moving-transform [ trans.mat, 1 ] --initialize-transforms-per-stage 0 --interpolation Linear --output [ output_, output_warped_image.nii.gz ] --transform Affine[ 2.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32, Random, 0.05 ] --convergence [ 1500x200, 1e-08, 20 ] --smoothing-sigmas 1.0x0.0vox --shrink-factors 2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --transform SyN[ 0.25, 3.0, 0.0 ] --metric Mattes[ fixed1.nii, moving1.nii, 1, 32 ] --convergence [ 100x50x30, 1e-09, 20 ] --smoothing-sigmas 2.0x1.0x0.0vox --shrink-factors 3x2x1 --use-estimate-learning-rate-once 1 --use-histogram-matching 1 --winsorize-image-intensities [ 0.0, 1.0 ]  --write-composite-transform 1'
+
     >>> # Test collapse transforms flag
     >>> reg4 = copy.deepcopy(reg)
     >>> reg.inputs.save_state = 'trans.mat'

nipype/interfaces/slicer/registration/brainsfit.py

@@ -67,6 +67,7 @@ class BRAINSFitInputSpec(CommandLineInputSpec):
     NEVER_USE_THIS_FLAG_IT_IS_OUTDATED_02 = traits.Bool(desc="DO NOT USE THIS FLAG", argstr="--NEVER_USE_THIS_FLAG_IT_IS_OUTDATED_02 ")
     permitParameterVariation = InputMultiPath(traits.Int, desc="A bit vector to permit linear transform parameters to vary under optimization.  The vector order corresponds with transform parameters, and beyond the end ones fill in as a default.  For instance, you can choose to rotate only in x (pitch) with 1,0,0;  this is mostly for expert use in turning on and off individual degrees of freedom in rotation, translation or scaling without multiplying the number of transform representations; this trick is probably meaningless when tried with the general affine transform.", sep=",", argstr="--permitParameterVariation %s")
     costMetric = traits.Enum("MMI", "MSE", "NC", "MC", desc="The cost metric to be used during fitting. Defaults to MMI. Options are MMI (Mattes Mutual Information), MSE (Mean Square Error), NC (Normalized Correlation), MC (Match Cardinality for binary images)", argstr="--costMetric %s")
+    writeOutputTransformInFloat = traits.Bool(desc="By default, the output registration transforms (either the output composite transform or each transform component) are written to the disk in double precision. If this flag is ON, the output transforms will be written in single (float) precision. It is especially important if the output transform is a displacement field transform, or it is a composite transform that includes several displacement fields.", argstr="--writeOutputTransformInFloat ")
 
 
 class BRAINSFitOutputSpec(TraitedSpec):

nipype/interfaces/slicer/registration/tests/test_auto_BRAINSFit.py

@@ -146,6 +146,8 @@ def test_BRAINSFit_inputs():
     ),
     writeTransformOnFailure=dict(argstr='--writeTransformOnFailure ',
     ),
+    writeOutputTransformInFloat=dict(argstr='--writeOutputTransformInFloat ',
+    ),
     )
     inputs = BRAINSFit.input_spec()