ITK  4.1.0
Insight Segmentation and Registration Toolkit
Public Types | Public Member Functions | Protected Member Functions | Private Attributes
itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample > Class Template Reference

#include <itkExpectationMaximizationMixtureModelEstimator.h>

+ Inheritance diagram for itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >:
+ Collaboration diagram for itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >:

List of all members.

Public Types

typedef MembershipFunctionBase
< MeasurementVectorType
ComponentMembershipFunctionType
typedef
MixtureModelComponentBase
< TSample > 
ComponentType
typedef std::vector
< ComponentType * > 
ComponentVectorType
typedef SmartPointer< const SelfConstPointer
typedef
GaussianMembershipFunctionType::Pointer 
GaussianMembershipFunctionPointer
typedef
GaussianMembershipFunction
< MeasurementVectorType
GaussianMembershipFunctionType
typedef TSample::MeasurementType MeasurementType
typedef
TSample::MeasurementVectorType 
MeasurementVectorType
typedef
MembershipFunctionType::ConstPointer 
MembershipFunctionPointer
typedef
SimpleDataObjectDecorator
< ProportionVectorType
MembershipFunctionsWeightsArrayObjectType
typedef
MembershipFunctionsWeightsArrayObjectType::Pointer 
MembershipFunctionsWeightsArrayPointer
typedef MembershipFunctionBase
< MeasurementVectorType
MembershipFunctionType
typedef
MembershipFunctionVectorObjectType::Pointer 
MembershipFunctionVectorObjectPointer
typedef
SimpleDataObjectDecorator
< MembershipFunctionVectorType
MembershipFunctionVectorObjectType
typedef std::vector
< MembershipFunctionPointer
MembershipFunctionVectorType
typedef SmartPointer< SelfPointer
typedef Array< double > ProportionVectorType
typedef TSample SampleType
typedef
ExpectationMaximizationMixtureModelEstimator 
Self
typedef Object Superclass
enum  TERMINATION_CODE {
  CONVERGED = 0,
  NOT_CONVERGED = 1
}

Public Member Functions

int AddComponent (ComponentType *component)
ComponentMembershipFunctionTypeGetComponentMembershipFunction (int componentIndex) const
int GetCurrentIteration ()
const ProportionVectorTypeGetInitialProportions () const
int GetMaximumIteration () const
const
MembershipFunctionsWeightsArrayObjectType
GetMembershipFunctionsWeightsArray () const
unsigned int GetNumberOfComponents () const
const
MembershipFunctionVectorObjectType
GetOutput () const
const ProportionVectorTypeGetProportions () const
const TSample * GetSample () const
TERMINATION_CODE GetTerminationCode () const
void SetInitialProportions (ProportionVectorType &propotion)
void SetMaximumIteration (int numberOfIterations)
void SetSample (const TSample *sample)
void Update ()

Protected Member Functions

bool CalculateDensities ()
double CalculateExpectation () const
 ExpectationMaximizationMixtureModelEstimator ()
void GenerateData ()
void PrintSelf (std::ostream &os, Indent indent) const
bool UpdateComponentParameters ()
bool UpdateProportions ()
virtual ~ExpectationMaximizationMixtureModelEstimator ()

Private Attributes

ComponentVectorType m_ComponentVector
int m_CurrentIteration
ProportionVectorType m_InitialProportions
int m_MaxIteration
MembershipFunctionVectorObjectPointer m_MembershipFunctionsObject
MembershipFunctionsWeightsArrayPointer m_MembershipFunctionsWeightArrayObject
ProportionVectorType m_Proportions
const TSample * m_Sample
TERMINATION_CODE m_TerminationCode
virtual const char * GetNameOfClass () const
virtual ::itk::LightObject::Pointer CreateAnother (void) const
static Pointer New ()

Detailed Description

template<class TSample>
class itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >

This class generates the parameter estimates for a mixture model using expectation maximization strategy.

The first template argument is the type of the target sample data. This estimator expects one or more mixture model component objects of the classes derived from the MixtureModelComponentBase. The actual component (or module) parameters are updated by each component. Users can think this class as a strategy or a integration point for the EM procedure. The initial proportion (SetInitialProportions), the input sample (SetSample), the mixture model components (AddComponent), and the maximum iteration (SetMaximumIteration) are required. The EM procedure terminates when the current iteration reaches the maximum iteration or the model parameters converge.

Recent API changes: The static const macro to get the length of a measurement vector, MeasurementVectorSize has been removed to allow the length of a measurement vector to be specified at run time. It is now obtained at run time from the sample set as input. Please use the function GetMeasurementVectorSize() to get the length.

See also:
MixtureModelComponentBase, GaussianMixtureModelComponent
Wiki Examples:

Definition at line 61 of file itkExpectationMaximizationMixtureModelEstimator.h.


Member Typedef Documentation

Type of the membership function base class

Definition at line 105 of file itkExpectationMaximizationMixtureModelEstimator.h.

Type of the mixture model component base class

Definition at line 98 of file itkExpectationMaximizationMixtureModelEstimator.h.

template<class TSample >
typedef std::vector< ComponentType * > itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::ComponentVectorType

Type of the component pointer storage

Definition at line 101 of file itkExpectationMaximizationMixtureModelEstimator.h.

Reimplemented from itk::Object.

Definition at line 68 of file itkExpectationMaximizationMixtureModelEstimator.h.

Typedef requried to generate dataobject decorated output that can be plugged into SampleClassifierFilter

Definition at line 84 of file itkExpectationMaximizationMixtureModelEstimator.h.

template<class TSample >
typedef TSample::MeasurementType itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::MeasurementType
template<class TSample >
typedef TSample::MeasurementVectorType itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::MeasurementVectorType

typedef for decorated array of proportion

Definition at line 127 of file itkExpectationMaximizationMixtureModelEstimator.h.

Reimplemented from itk::Object.

Definition at line 67 of file itkExpectationMaximizationMixtureModelEstimator.h.

template<class TSample >
typedef Array< double > itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::ProportionVectorType

Type of the array of the proportion values

Definition at line 108 of file itkExpectationMaximizationMixtureModelEstimator.h.

template<class TSample >
typedef TSample itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::SampleType

TSample template argument related typedefs

Definition at line 73 of file itkExpectationMaximizationMixtureModelEstimator.h.

Standard class typedef

Reimplemented from itk::Object.

Definition at line 65 of file itkExpectationMaximizationMixtureModelEstimator.h.

Reimplemented from itk::Object.

Definition at line 66 of file itkExpectationMaximizationMixtureModelEstimator.h.


Member Enumeration Documentation

Termination status after running optimization

Enumerator:
CONVERGED 
NOT_CONVERGED 

Definition at line 158 of file itkExpectationMaximizationMixtureModelEstimator.h.


Constructor & Destructor Documentation

template<class TSample >
virtual itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::~ExpectationMaximizationMixtureModelEstimator ( ) [inline, protected, virtual]

Member Function Documentation

template<class TSample >
int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::AddComponent ( ComponentType component)

Adds a new component (or class).

template<class TSample >
bool itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::CalculateDensities ( ) [protected]
template<class TSample >
double itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::CalculateExpectation ( ) const [protected]
template<class TSample >
virtual::itk::LightObject::Pointer itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::CreateAnother ( void  ) const [virtual]

Standard macros

Reimplemented from itk::Object.

template<class TSample >
void itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GenerateData ( ) [protected]

Starts the estimation process

Gets the membership function specified by componentIndex argument.

template<class TSample >
int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GetCurrentIteration ( ) [inline]

Gets the current iteration.

Definition at line 143 of file itkExpectationMaximizationMixtureModelEstimator.h.

template<class TSample >
int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GetMaximumIteration ( ) const

Get method for data decorated Membership functions weights array

template<class TSample >
virtual const char* itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GetNameOfClass ( ) const [virtual]

Standard macros

Reimplemented from itk::Object.

template<class TSample >
unsigned int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GetNumberOfComponents ( ) const

Gets the total number of classes currently plugged in.

Output Membership function vector containing the membership functions with the final optimized parameters

Gets the result proportion values

template<class TSample >
const TSample* itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::GetSample ( ) const

Returns the target data

Gets the termination status

template<class TSample >
static Pointer itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::New ( ) [static]

Standard macros

Reimplemented from itk::Object.

template<class TSample >
void itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::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::Object.

Set/Gets the initial proportion values. The size of proportion vector should be same as the number of component (or classes)

template<class TSample >
void itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::SetMaximumIteration ( int  numberOfIterations)

Set/Gets the maximum number of iterations. When the optimization process reaches the maximum number of interations, even if the class parameters aren't converged, the optimization process stops.

template<class TSample >
void itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::SetSample ( const TSample *  sample)

Sets the target data that will be classified by this

template<class TSample >
void itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::Update ( )

Runs the optimization process.

template<class TSample >
bool itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::UpdateComponentParameters ( ) [protected]
template<class TSample >
bool itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::UpdateProportions ( ) [protected]

Member Data Documentation

template<class TSample >
int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::m_CurrentIteration [private]
template<class TSample >
int itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::m_MaxIteration [private]
template<class TSample >
const TSample* itk::Statistics::ExpectationMaximizationMixtureModelEstimator< TSample >::m_Sample [private]

Target data sample pointer

Definition at line 189 of file itkExpectationMaximizationMixtureModelEstimator.h.


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