ITK  4.1.0
Insight Segmentation and Registration Toolkit
Public Types | Public Member Functions | Static Public Member Functions | Protected Member Functions | Private Member Functions
itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField > Class Template Reference

#include <itkLevelSetMotionRegistrationFilter.h>

+ Inheritance diagram for itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >:
+ Collaboration diagram for itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >:

List of all members.

Public Types

typedef SmartPointer< const SelfConstPointer
typedef
Superclass::DisplacementFieldPointer 
DisplacementFieldPointer
typedef
Superclass::DisplacementFieldType 
DisplacementFieldType
typedef
Superclass::FiniteDifferenceFunctionType 
FiniteDifferenceFunctionType
typedef
Superclass::FixedImagePointer 
FixedImagePointer
typedef Superclass::FixedImageType FixedImageType
typedef
LevelSetMotionRegistrationFunction
< FixedImageType,
MovingImageType,
DisplacementFieldType
LevelSetMotionFunctionType
typedef
Superclass::MovingImagePointer 
MovingImagePointer
typedef Superclass::MovingImageType MovingImageType
typedef SmartPointer< SelfPointer
typedef
LevelSetMotionRegistrationFilter 
Self
typedef
PDEDeformableRegistrationFilter
< TFixedImage, TMovingImage,
TDisplacementField > 
Superclass
typedef Superclass::TimeStepType TimeStepType

Public Member Functions

virtual ::itk::LightObject::Pointer CreateAnother (void) const
virtual double GetAlpha () const
virtual double GetGradientMagnitudeThreshold () const
virtual double GetGradientSmoothingStandardDeviations () const
virtual double GetIntensityDifferenceThreshold () const
virtual double GetMetric () const
virtual const char * GetNameOfClass () const
virtual void SetAlpha (double)
virtual void SetGradientMagnitudeThreshold (double)
virtual void SetGradientSmoothingStandardDeviations (double sigma)
virtual void SetIntensityDifferenceThreshold (double)

Static Public Member Functions

static Pointer New ()

Protected Member Functions

virtual void ApplyUpdate (const TimeStepType &dt)
virtual bool Halt ()
virtual void InitializeIteration ()
 LevelSetMotionRegistrationFilter ()
void PrintSelf (std::ostream &os, Indent indent) const
 ~LevelSetMotionRegistrationFilter ()

Private Member Functions

 LevelSetMotionRegistrationFilter (const Self &)
void operator= (const Self &)

Detailed Description

template<class TFixedImage, class TMovingImage, class TDisplacementField>
class itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >

Deformably register two images using level set motion.

LevelSetMotionFilter implements a deformable registration algorithm that aligns a fixed and a moving image under level set motion. The equations of motion are similar to those of the DemonsRegistrationFilter. The main differences are: (1) Gradients of the moving image are calculated on a smoothed image while intensity difference are measured on the original images (2) Magnitude of the motion vector is a function of the differences in intensity between the fixed and moving pixel. An adaptive timestep is calculated based on the maximum motion vector over the entire field to ensure stability. The timestep also implictly converts the motion vector measured in units of intensity to a vector measured in physical units. Demons, on the other hand, defines its motion vectors as function of both the intensity differences and gradient magnitude at each respective pixel. Consider two separate pixels with the same intensity differences between the corresponding fixed and moving pixel pairs. In demons, the motion vector of the pixel over a low gradient region will be larger than the motion vector of the pixel over a large gradient region. This leads to an unstable vector field. In the levelset approach, the motion vectors will be proportional to the gradients, scaled by the maximum gradient over the entire field. The pixel with at the lower gradient position will more less than the pixel at the higher gradient position. (3) Gradients are calculated using minmod finite difference instead of using central differences.

A deformation field is represented as a image whose pixel type is some vector type with at least N elements, where N is the dimension of the fixed image. The vector type must support element access via operator []. It is assumed that the vector elements behave like floating point scalars.

This class is templated over the fixed image type, moving image type and the deformation field type.

The input fixed and moving images are set via methods SetFixedImage and SetMovingImage respectively. An initial deformation field maybe set via SetInitialDisplacementField or SetInput. If no initial field is set, a zero field is used as the initial condition.

The algorithm has one parameters: the number of iteration to be performed.

The output deformation field can be obtained via methods GetOutput or GetDisplacementField.

This class make use of the finite difference solver hierarchy. Update for each iteration is computed in LevelSetMotionFunction.

Warning:
This filter assumes that the fixed image type, moving image type and deformation field type all have the same number of dimensions.

Ref: B.C. Vemuri, J. Ye, Y. Chen, C.M. Leonard. "Image registration via level-set motion: applications to atlas-based segmentation". Medical Image Analysis. Vol. 7. pp. 1-20. 2003.

See also:
LevelSetMotionRegistrationFunction
DemonsRegistrationFilter

Definition at line 91 of file itkLevelSetMotionRegistrationFilter.h.


Member Typedef Documentation

template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef SmartPointer< const Self > itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::ConstPointer
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::DisplacementFieldPointer itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::DisplacementFieldPointer
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::DisplacementFieldType itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::DisplacementFieldType
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::FiniteDifferenceFunctionType itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::FiniteDifferenceFunctionType
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::FixedImagePointer itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::FixedImagePointer
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::FixedImageType itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::FixedImageType
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef LevelSetMotionRegistrationFunction< FixedImageType, MovingImageType, DisplacementFieldType > itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::LevelSetMotionFunctionType

LevelSetMotionFilterFunction type.

Definition at line 132 of file itkLevelSetMotionRegistrationFilter.h.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::MovingImagePointer itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::MovingImagePointer
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::MovingImageType itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::MovingImageType
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef SmartPointer< Self > itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::Pointer
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef LevelSetMotionRegistrationFilter itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::Self
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField > itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::Superclass
template<class TFixedImage , class TMovingImage , class TDisplacementField >
typedef Superclass::TimeStepType itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::TimeStepType

Inherit types from superclass.

Reimplemented from itk::DenseFiniteDifferenceImageFilter< TDisplacementField, TDisplacementField >.

Definition at line 107 of file itkLevelSetMotionRegistrationFilter.h.


Constructor & Destructor Documentation

template<class TFixedImage , class TMovingImage , class TDisplacementField >
itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::LevelSetMotionRegistrationFilter ( ) [protected]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::~LevelSetMotionRegistrationFilter ( ) [inline, protected]

Definition at line 183 of file itkLevelSetMotionRegistrationFilter.h.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::LevelSetMotionRegistrationFilter ( const Self ) [private]

Member Function Documentation

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::ApplyUpdate ( const TimeStepType dt) [protected, virtual]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual::itk::LightObject::Pointer itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::CreateAnother ( void  ) const [virtual]

Create an object from an instance, potentially deferring to a factory. This method allows you to create an instance of an object that is exactly the same type as the referring object. This is useful in cases where an object has been cast back to a base class.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual double itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetAlpha ( ) const [virtual]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual double itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetGradientMagnitudeThreshold ( ) const [virtual]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual double itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetGradientSmoothingStandardDeviations ( ) const [virtual]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual double itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetIntensityDifferenceThreshold ( ) const [virtual]
template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual double itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetMetric ( ) const [virtual]

Get the metric value. The metric value is the mean square difference in intensity between the fixed image and transforming moving image computed over the the overlapping region between the two images. This is value is only available for the previous iteration and NOT the current iteration.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual const char* itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::GetNameOfClass ( ) const [virtual]

Run-time type information (and related methods).

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual bool itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::Halt ( ) [protected, virtual]

This method returns true when the current iterative solution of the equation has met the criteria to stop solving. This version calls the superclass' version but also Halts if the RMSChange is zero.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::InitializeIteration ( ) [protected, virtual]

Initialize the state of filter and equation before each iteration.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
static Pointer itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::New ( ) [static]

Method for creation through the object factory.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::operator= ( const Self ) [private]

PushBackInput(), PushFronInput() in the public section force the input to be the type expected by an ImageToImageFilter. However, these methods end of "hiding" the versions from the superclass (ProcessObject) whose arguments are DataObjects. Here, we re-expose the versions from ProcessObject to avoid warnings about hiding methods from the superclass.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::PrintSelf ( std::ostream &  os,
Indent  indent 
) const [protected, virtual]

Set/Get the desired limits of the Gaussian kernel width.

See also:
GaussianOperator.

Reimplemented from itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::SetAlpha ( double  ) [virtual]

Set/Get the parameter alpha. Alpha is added to the calculated gradient magnitude prior to normalizing the gradient to protect against numerical instability as the gradient magnitude approaches zero. This should be set as a small fraction of the intensity dynamic range, for instance 0.04%. Default is the absolute (not percentage) value of 0.1.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::SetGradientMagnitudeThreshold ( double  ) [virtual]

Set/Get the threshold below which the gradient magnitude is considered the zero vector. Default is 1e-9.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::SetGradientSmoothingStandardDeviations ( double  sigma) [virtual]

Set/Get the standard deviation used for smoothing the moving image prior to calculating gradients. The standard deviation is measured in physical units (for instance mm). Note that this smoothing value is not to be confused with the PDEDeformableRegistrationFilter::SetStandardDeviations() method. The method in PDEDeformableRegistrationFilter is for setting the smoothing parameters for regularizing the deformation field between interations. Those smoothing parameters are set in pixel units not physical units. Deformation field smoothing is not done by default in LevelSetMotionRegistration. This smoothing parameter is to condition the gradient calculation and parameter is specified in physical units.

template<class TFixedImage , class TMovingImage , class TDisplacementField >
virtual void itk::LevelSetMotionRegistrationFilter< TFixedImage, TMovingImage, TDisplacementField >::SetIntensityDifferenceThreshold ( double  ) [virtual]

Set/Get the threshold below which the absolute difference of intensity yields a match. When the intensities match between a moving and fixed image pixel, the update vector (for that iteration) will be the zero vector. Default is 0.001.


The documentation for this class was generated from the following file: