ITK  4.1.0
Insight Segmentation and Registration Toolkit
Public Types | Public Member Functions | Static Public Member Functions | Protected Member Functions | Private Member Functions | Private Attributes
itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage > Class Template Reference

#include <itkRecursiveGaussianImageFilter.h>

+ Inheritance diagram for itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >:
+ Collaboration diagram for itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >:

List of all members.

Public Types

typedef SmartPointer< const SelfConstPointer
enum  OrderEnumType {
  ZeroOrder,
  FirstOrder,
  SecondOrder
}
typedef TOutputImage OutputImageType
typedef SmartPointer< SelfPointer
typedef Superclass::RealType RealType
typedef Superclass::ScalarRealType ScalarRealType
typedef
RecursiveGaussianImageFilter 
Self
typedef
RecursiveSeparableImageFilter
< TInputImage, TOutputImage > 
Superclass

Public Member Functions

virtual ::itk::LightObject::Pointer CreateAnother (void) const
virtual const char * GetNameOfClass () const
virtual bool GetNormalizeAcrossScale () const
void SetFirstOrder ()
virtual void SetNormalizeAcrossScale (bool _arg)
void SetSecondOrder ()
void SetZeroOrder ()
virtual ScalarRealType GetSigma () const
virtual void SetSigma (ScalarRealType _arg)
virtual void SetOrder (OrderEnumType _arg)
virtual OrderEnumType GetOrder () const

Static Public Member Functions

static Pointer New ()

Protected Member Functions

void PrintSelf (std::ostream &os, Indent indent) const
 RecursiveGaussianImageFilter ()
virtual void SetUp (ScalarRealType spacing)
virtual ~RecursiveGaussianImageFilter ()

Private Member Functions

void ComputeDCoefficients (ScalarRealType sigmad, ScalarRealType W1, ScalarRealType L1, ScalarRealType W2, ScalarRealType L2, ScalarRealType &SD, ScalarRealType &DD, ScalarRealType &ED)
void ComputeNCoefficients (ScalarRealType sigmad, ScalarRealType A1, ScalarRealType B1, ScalarRealType W1, ScalarRealType L1, ScalarRealType A2, ScalarRealType B2, ScalarRealType W2, ScalarRealType L2, ScalarRealType &N0, ScalarRealType &N1, ScalarRealType &N2, ScalarRealType &N3, ScalarRealType &SN, ScalarRealType &DN, ScalarRealType &EN)
void ComputeRemainingCoefficients (bool symmetric)
void operator= (const Self &)
 RecursiveGaussianImageFilter (const Self &)

Private Attributes

bool m_NormalizeAcrossScale
OrderEnumType m_Order
ScalarRealType m_Sigma

Detailed Description

template<typename TInputImage, typename TOutputImage = TInputImage>
class itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >

Base class for computing IIR convolution with an approximation of a Gaussian kernel.

\[ \frac{ 1 }{ \sigma \sqrt{ 2 \pi } } \exp{ \left( - \frac{x^2}{ 2 \sigma^2 } \right) } \]

RecursiveGaussianImageFilter is the base class for recursive filters that approximate convolution with the Gaussian kernel. This class implements the recursive filtering method proposed by R.Deriche in IEEE-PAMI Vol.12, No.1, January 1990, pp 78-87, "Fast Algorithms for Low-Level Vision"

Details of the implementation are described in the technical report: R. Deriche, "Recursively Implementing The Gaussian and Its Derivatives", INRIA, 1993, ftp://ftp.inria.fr/INRIA/tech-reports/RR/RR-1893.ps.gz

Further improvements of the algorithm are described in: G. Farneback & C.-F. Westin, "On Implementation of Recursive Gaussian Filters", so far unpublished.

As compared to itk::DiscreteGaussianImageFilter, this filter tends to be faster for large kernels, and it can take the derivative of the blurred image in one step. Also, note that we have itk::RecursiveGaussianImageFilter::SetSigma(), but itk::DiscreteGaussianImageFilter::SetVariance().

See also:
DiscreteGaussianImageFilter
Wiki Examples:

Definition at line 62 of file itkRecursiveGaussianImageFilter.h.


Member Typedef Documentation

template<typename TInputImage , typename TOutputImage = TInputImage>
typedef SmartPointer< const Self > itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::ConstPointer
template<typename TInputImage , typename TOutputImage = TInputImage>
typedef TOutputImage itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::OutputImageType

Type of the output image

Reimplemented from itk::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

Definition at line 93 of file itkRecursiveGaussianImageFilter.h.

template<typename TInputImage , typename TOutputImage = TInputImage>
typedef SmartPointer< Self > itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::Pointer
template<typename TInputImage , typename TOutputImage = TInputImage>
typedef Superclass::RealType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::RealType
template<typename TInputImage , typename TOutputImage = TInputImage>
typedef Superclass::ScalarRealType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::ScalarRealType
template<typename TInputImage , typename TOutputImage = TInputImage>
typedef RecursiveGaussianImageFilter itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::Self

Standard class typedefs.

Reimplemented from itk::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

Definition at line 67 of file itkRecursiveGaussianImageFilter.h.

template<typename TInputImage , typename TOutputImage = TInputImage>
typedef RecursiveSeparableImageFilter< TInputImage, TOutputImage > itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::Superclass

Member Enumeration Documentation

template<typename TInputImage , typename TOutputImage = TInputImage>
enum itk::RecursiveGaussianImageFilter::OrderEnumType

Enum type that indicates if the filter applies the equivalent operation of convolving with a gaussian, first derivative of a gaussian or the second derivative of a gaussian.

Enumerator:
ZeroOrder 
FirstOrder 
SecondOrder 

Definition at line 90 of file itkRecursiveGaussianImageFilter.h.


Constructor & Destructor Documentation

template<typename TInputImage , typename TOutputImage = TInputImage>
itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::RecursiveGaussianImageFilter ( ) [protected]
template<typename TInputImage , typename TOutputImage = TInputImage>
virtual itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::~RecursiveGaussianImageFilter ( ) [inline, protected, virtual]

Definition at line 158 of file itkRecursiveGaussianImageFilter.h.

template<typename TInputImage , typename TOutputImage = TInputImage>
itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::RecursiveGaussianImageFilter ( const Self ) [private]

Member Function Documentation

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::ComputeDCoefficients ( ScalarRealType  sigmad,
ScalarRealType  W1,
ScalarRealType  L1,
ScalarRealType  W2,
ScalarRealType  L2,
ScalarRealType SD,
ScalarRealType DD,
ScalarRealType ED 
) [private]

Compute the D coefficients in the recursive filter.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::ComputeNCoefficients ( ScalarRealType  sigmad,
ScalarRealType  A1,
ScalarRealType  B1,
ScalarRealType  W1,
ScalarRealType  L1,
ScalarRealType  A2,
ScalarRealType  B2,
ScalarRealType  W2,
ScalarRealType  L2,
ScalarRealType N0,
ScalarRealType N1,
ScalarRealType N2,
ScalarRealType N3,
ScalarRealType SN,
ScalarRealType DN,
ScalarRealType EN 
) [private]

Compute the N coefficients in the recursive filter.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::ComputeRemainingCoefficients ( bool  symmetric) [private]

Compute the M coefficients and the boundary coefficients in the recursive filter.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual::itk::LightObject::Pointer itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::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::Object.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual const char* itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::GetNameOfClass ( ) const [virtual]

Type macro that defines a name for this class

Reimplemented from itk::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual bool itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::GetNormalizeAcrossScale ( ) const [virtual]
template<typename TInputImage , typename TOutputImage = TInputImage>
virtual OrderEnumType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::GetOrder ( ) const [virtual]

Set/Get the Order of the Gaussian to convolve with.

  • ZeroOrder is equivalent to convolving with a Gaussian. This is the default.
  • FirstOrder is equivalent to convolving with the first derivative of a Gaussian.
  • SecondOrder is equivalent to convolving with the second derivative of a Gaussian.
template<typename TInputImage , typename TOutputImage = TInputImage>
virtual ScalarRealType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::GetSigma ( ) const [virtual]

Set/Get the Sigma, measured in world coordinates, of the Gaussian kernel. The default is 1.0.

template<typename TInputImage , typename TOutputImage = TInputImage>
static Pointer itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::New ( ) [static]

Method for creation through the object factory.

Reimplemented from itk::Object.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::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::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::PrintSelf ( std::ostream &  os,
Indent  indent 
) const [protected, virtual]

Methods invoked by Print() to print information about the object including superclasses. Typically not called by the user (use Print() instead) but used in the hierarchical print process to combine the output of several classes.

Reimplemented from itk::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetFirstOrder ( )

Explicitly set a first order derivative.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetNormalizeAcrossScale ( bool  _arg) [virtual]

Set/Get the flag for normalizing the gaussian over scale-space.

This flag enables the analysis of the differential shape of features independent of their size ( both pixels and physical size ). Following the notation of Tony Lindeberg:

Let

\[ L(x; t) = g(x; t) \ast f(x) \]

be the scale-space representation of image

\[ f(x) \]

where

\[ g(x; t) = \frac{1}{ \sqrt{ 2 \pi t} } \exp{ \left( -\frac{x^2}{ 2 t } \right) } \]

is the Gaussian function and

\[\ast\]

denotes convolution. This is a change from above with

\[ t = \sigma^2 \]

.

Then the normalized derivative operator for normalized coordinates across scale is:

\[ \partial_\xi = \sqrt{t} \partial_x \]

The resulting scaling factor is

\[ \sigma^N \]

where N is the order of the derivative.

When this flag is ON the filter will be normalized in such a way that the values of derivatives are not biased by the size of the object. That is to say the maximum value a feature reaches across scale is independent of the scale of the object.

For analyzing an image across scale-space you want to enable this flag. It is disabled by default.

Note:
Not all scale space axioms are satisfied by this filter, some are only approximated. Particularly, at fine scales ( say less than 1 pixel ) other methods such as a discrete Gaussian kernel should be considered.
template<typename TInputImage , typename TOutputImage = TInputImage>
virtual void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetOrder ( OrderEnumType  _arg) [virtual]

Set/Get the Order of the Gaussian to convolve with.

  • ZeroOrder is equivalent to convolving with a Gaussian. This is the default.
  • FirstOrder is equivalent to convolving with the first derivative of a Gaussian.
  • SecondOrder is equivalent to convolving with the second derivative of a Gaussian.
template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetSecondOrder ( )

Explicitly set a second order derivative.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetSigma ( ScalarRealType  _arg) [virtual]

Set/Get the Sigma, measured in world coordinates, of the Gaussian kernel. The default is 1.0.

template<typename TInputImage , typename TOutputImage = TInputImage>
virtual void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetUp ( ScalarRealType  spacing) [protected, virtual]

Set up the coefficients of the filter to approximate a specific kernel. Here it is used to approximate a Gaussian or one of its derivatives. Parameter is the spacing along the dimension to filter.

Implements itk::RecursiveSeparableImageFilter< TInputImage, TOutputImage >.

template<typename TInputImage , typename TOutputImage = TInputImage>
void itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::SetZeroOrder ( )

Explicitly set a zeroth order derivative.


Member Data Documentation

template<typename TInputImage , typename TOutputImage = TInputImage>
bool itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::m_NormalizeAcrossScale [private]

Normalize the image across scale space

Definition at line 192 of file itkRecursiveGaussianImageFilter.h.

template<typename TInputImage , typename TOutputImage = TInputImage>
OrderEnumType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::m_Order [private]

Definition at line 194 of file itkRecursiveGaussianImageFilter.h.

template<typename TInputImage , typename TOutputImage = TInputImage>
ScalarRealType itk::RecursiveGaussianImageFilter< TInputImage, TOutputImage >::m_Sigma [private]

Sigma of the gaussian kernel.

Definition at line 189 of file itkRecursiveGaussianImageFilter.h.


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