Main Page Groups Namespace List Class Hierarchy Alphabetical List Compound List File List Namespace Members Compound Members File Members Concepts

itk::ProcessObject Class Reference
[ITK System Objects, Data Processing Objects]

ProcessObject is the base class for all process objects (source, filters, mappers) in the Insight data processing pipeline. More...

#include <itkProcessObject.h>

Inheritance diagram for itk::ProcessObject:

[legend]Collaboration diagram for itk::ProcessObject:

[legend]List of all members.


Public Types
typedef ProcessObject	Self
typedef Object	Superclass
typedef SmartPointer< Self >	Pointer
typedef SmartPointer< const Self >	ConstPointer
typedef DataObject::Pointer	DataObjectPointer
typedef std::vector< DataObjectPointer >	DataObjectPointerArray
Public Member Functions
virtual const char *	GetNameOfClass () const
DataObjectPointerArray &	GetInputs ()
std::vector< DataObjectPointer >::size_type	GetNumberOfInputs () const
virtual std::vector< DataObjectPointer >::size_type	GetNumberOfValidRequiredInputs () const
virtual void	SetAbortGenerateData (bool _arg)
virtual const bool &	GetAbortGenerateData ()
virtual void	AbortGenerateDataOn ()
virtual void	AbortGenerateDataOff ()
virtual void	SetProgress (float _arg)
virtual const float &	GetProgress ()
void	UpdateProgress (float amount)
virtual void	Update ()
virtual void	UpdateLargestPossibleRegion ()
virtual void	UpdateOutputInformation ()
virtual void	PropagateRequestedRegion (DataObject *output)
virtual void	UpdateOutputData (DataObject *output)
virtual void	EnlargeOutputRequestedRegion (DataObject *)
virtual void	ResetPipeline ()
virtual DataObjectPointer	MakeOutput (unsigned int idx)
MultiThreader *	GetMultiThreader ()
virtual void	PrepareOutputs ()

DataObjectPointerArray &	GetOutputs ()
std::vector< DataObjectPointer >::size_type	GetNumberOfOutputs () const

virtual void	SetReleaseDataFlag (bool flag)
virtual bool	GetReleaseDataFlag () const
void	ReleaseDataFlagOn ()
void	ReleaseDataFlagOff ()

virtual void	SetReleaseDataBeforeUpdateFlag (bool _arg)
virtual const bool &	GetReleaseDataBeforeUpdateFlag ()
virtual void	ReleaseDataBeforeUpdateFlagOn ()
virtual void	ReleaseDataBeforeUpdateFlagOff ()

virtual void	SetNumberOfThreads (int _arg)
virtual const int &	GetNumberOfThreads ()
Protected Member Functions
	ProcessObject ()
	~ProcessObject ()
void	PrintSelf (std::ostream &os, Indent indent) const
void	SetNumberOfInputs (unsigned int num)
void	SetNumberOfOutputs (unsigned int num)
virtual void	GenerateInputRequestedRegion ()
virtual void	GenerateOutputRequestedRegion (DataObject *output)
virtual void	GenerateOutputInformation ()
virtual void	GenerateData ()
virtual void	PropagateResetPipeline ()
virtual void	ReleaseInputs ()
virtual void	CacheInputReleaseDataFlags ()
virtual void	RestoreInputReleaseDataFlags ()

virtual void	SetNthInput (unsigned int num, DataObject *input)
virtual void	AddInput (DataObject *input)
virtual void	RemoveInput (DataObject *input)
virtual void	SetNumberOfRequiredInputs (unsigned int _arg)
virtual const unsigned int &	GetNumberOfRequiredInputs ()

DataObject *	GetInput (unsigned int idx)
const DataObject *	GetInput (unsigned int idx) const

virtual void	SetNthOutput (unsigned int num, DataObject *output)
virtual void	AddOutput (DataObject *output)
virtual void	RemoveOutput (DataObject *output)
virtual void	SetNumberOfRequiredOutputs (unsigned int _arg)
virtual const unsigned int &	GetNumberOfRequiredOutputs ()

DataObject *	GetOutput (unsigned int idx)
const DataObject *	GetOutput (unsigned int idx) const
Protected Attributes
bool	m_Updating
TimeStamp	m_OutputInformationMTime
Friends
class	DataObject

Detailed Description

ProcessObject is the base class for all process objects (source, filters, mappers) in the Insight data processing pipeline.

ProcessObject is an abstract object that specifies behavior and interface of network process objects (sources, filters, mappers). Source objects are creators of visualization data; filters input, process, and output image data; and mappers transform data into another form (like transforming coordinates or writing data to a file).

A major role of ProcessObject is to define the inputs and outputs of a filter. More than one input and/or output may exist for a given filter. Some classes (e.g., source objects or mapper objects) will not use inputs (the source) or outputs (mappers). In this case, the inputs or outputs is just ignored.

ProcessObject invokes the following events: Command::StartEvent, Command::EndEvent These are convenience events you can use for any purpose (e.g., debugging info, highlighting/notifying user interface, etc.) See Command and LightObject for information on using AddObserver.

Another event Command::ProgressEvent can be observed. Some filters invoke this event periodically during their execution (with the progress, parameter, the fraction of work done). The use is similar to that of StartEvent and EndEvent. Filters may also check their AbortGenerateData flag to determine whether to prematurally end their execution.

An important feature of subclasses of ProcessObject is that it is possible to control the memory-management model (i.e., retain output versus delete output data). The ReleaseDataFlag enables the deletion of the output data once the downstream process object finishes processing the data (please see text). The ReleaseDataBeforeUpdateFlag enables the deletion of the ProcessObject's output data from a previous update if that output data is slated to be regenerated by the pipeline process. Setting this flag can control peak memory usage during a subsequent pipeline update. For a ProcessObject, the ReleaseDataFlag defaults to false and the ReleaseDataBeforeUpdateFlag defaults to true. Some subclasses of ProcessObject, for example ImageSource, use a default setting of false for the ReleaseDataBeforeUpdateFlag.

Subclasses of ProcessObject may override 4 of the methods of this class to control how a given filter may interact with the pipeline (dataflow). These methods are: GenerateOutputInformation(), EnlargeOutputRequestedRegion(), GenerateInputRequestedRegion(), and GenerateOutputRequestedRegion(). By overriding these methods, a filter can deviate from the base assumptions of the pipeline execution model.

Definition at line 87 of file itkProcessObject.h.

Member Typedef Documentation

typedef std::vector<DataObjectPointer> itk::ProcessObject::DataObjectPointerArray

STL Array of SmartPointers to DataObjects
Definition at line 103 of file itkProcessObject.h.
Referenced by GetInputs(), and GetOutputs().

Constructor & Destructor Documentation

itk::ProcessObject::ProcessObject ( ) [protected]

itk::ProcessObject::~ProcessObject ( ) [protected]

Member Function Documentation

virtual void itk::ProcessObject::AbortGenerateDataOff ( ) [virtual]

virtual void itk::ProcessObject::AbortGenerateDataOn ( ) [virtual]

Turn on and off the AbortGenerateData flag.

virtual void itk::ProcessObject::AddInput ( DataObject * input ) [protected, virtual]

Protected methods for setting inputs. Subclasses make use of them for setting input.

virtual void itk::ProcessObject::AddOutput ( DataObject * output ) [protected, virtual]

Protected methods for setting outputs. Subclasses make use of them for getting output.

virtual void itk::ProcessObject::CacheInputReleaseDataFlags ( ) [protected, virtual]

Cache the state of any ReleaseDataFlag's on the inputs. While the filter is executing, we need to set the ReleaseDataFlag's on the inputs to false in case the current filter is implemented using a mini-pipeline (which will try to release the inputs). After the filter finishes, we restore the state of the ReleaseDataFlag's before the call to ReleaseInputs().

virtual void itk::ProcessObject::GenerateOutputRequestedRegion ( DataObject * output ) [protected, virtual]

Given one output whose requested region has been set, how should the requested regions for the remaining outputs of the process object be set? By default, all the outputs are set to the same requested region. If a filter needs to produce different requested regions for each output, for instance an image processing filter producing several outputs at different resolutions, then that filter may override this method and set the requested regions appropriatedly.
Note that a filter producing multiple outputs of different types is required to override this method. The default implementation can only correctly handle multiple outputs of the same type.
Reimplemented in itk::MultiResolutionPyramidImageFilter< TInputImage, TOutputImage >, itk::RecursiveMultiResolutionPyramidImageFilter< TInputImage, TOutputImage >, itk::watershed::BoundaryResolver< TPixelType, TDimension >, itk::watershed::EquivalenceRelabeler< TScalarType, TImageDimension >, itk::watershed::Relabeler< TScalarType, TImageDimension >, itk::watershed::Segmenter< TInputImage >, itk::watershed::SegmentTreeGenerator< TScalarType >, itk::watershed::Relabeler< ScalarType, itkGetStaticConstMacro(ImageDimension)>, itk::watershed::Segmenter< InputImageType >, and itk::watershed::SegmentTreeGenerator< ScalarType >.

virtual const bool& itk::ProcessObject::GetAbortGenerateData ( ) [virtual]

Get the AbortGenerateData flag for the process object. Process objects may handle premature termination of execution in different ways.

const DataObject* itk::ProcessObject::GetInput ( unsigned int idx ) const [protected]

Method used internally for getting an input.

DataObjectPointerArray& itk::ProcessObject::GetInputs ( ) [inline]

Return an array with all the inputs of this process object. This is useful for tracing back in the pipeline to construct graphs etc.
Definition at line 108 of file itkProcessObject.h.
References DataObjectPointerArray.

MultiThreader* itk::ProcessObject::GetMultiThreader ( ) [inline]

Return the multithreader used by this class.
Definition at line 281 of file itkProcessObject.h.

std::vector<DataObjectPointer>::size_type itk::ProcessObject::GetNumberOfInputs ( ) const [inline]

Get the size of the input vector. This is merely the size of the input vector, not the number of inputs that have valid DataObject's assigned. Use GetNumberOfValidRequiredInputs() to determine how many inputs are non-null.
Definition at line 115 of file itkProcessObject.h.

std::vector<DataObjectPointer>::size_type itk::ProcessObject::GetNumberOfOutputs ( ) const [inline]

Return an array with all the outputs of this process object. This is useful for tracing forward in the pipeline to contruct graphs etc.
Definition at line 132 of file itkProcessObject.h.

virtual const unsigned int& itk::ProcessObject::GetNumberOfRequiredInputs ( ) [protected, virtual]

Protected methods for setting inputs. Subclasses make use of them for setting input.

virtual const unsigned int& itk::ProcessObject::GetNumberOfRequiredOutputs ( ) [protected, virtual]

Protected methods for setting outputs. Subclasses make use of them for getting output.

virtual const int& itk::ProcessObject::GetNumberOfThreads ( ) [virtual]

Get/Set the number of threads to create when executing.

virtual std::vector<DataObjectPointer>::size_type itk::ProcessObject::GetNumberOfValidRequiredInputs ( ) const [virtual]

Get the number of valid inputs. This is the number of non-null entries in the input vector in the first NumberOfRequiredInputs slots. This method is used to determine whether the necessary required inputs have been set. Subclasses of ProcessObject may override this implementation if the required inputs are not the first slots in input vector.
Reimplemented in itk::MultiResolutionPDEDeformableRegistration< TFixedImage, TMovingImage, TDeformationField >, and itk::PDEDeformableRegistrationFilter< TFixedImage, TMovingImage, TDeformationField >.

const DataObject* itk::ProcessObject::GetOutput ( unsigned int idx ) const [protected]

Method used internally for getting an output.

DataObjectPointerArray& itk::ProcessObject::GetOutputs ( ) [inline]

Return an array with all the outputs of this process object. This is useful for tracing forward in the pipeline to contruct graphs etc.
Definition at line 130 of file itkProcessObject.h.
References DataObjectPointerArray.

virtual const float& itk::ProcessObject::GetProgress ( ) [virtual]

Get the execution progress of a process object. The progress is a floating number in [0,1] with 0 meaning no progress and 1 meaning the filter has completed execution.

virtual const bool& itk::ProcessObject::GetReleaseDataBeforeUpdateFlag ( ) [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released/reallocated during an Update(). In limited memory scenarios, a user may want to force the elements of a pipeline to release any bulk data that is going to be regenerated anyway during an Update() in order to control peak memory allocation. Note that this flag is different from the ReleaseDataFlag. ReleaseDataFlag manages the deallocation of a ProcessObject's bulk output data once that data has been consumed by a downstream ProcessObject. The ReleaseDataBeforeUpdateFlag manages the deallocation/reallocation of bulk data during a pipeline update to control peak memory utilization. Default value is on.

virtual bool itk::ProcessObject::GetReleaseDataFlag ( ) const [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released after being used by a downstream ProcessObject. Default value is off. Another options for controlling memory utilization is the ReleaseDataBeforeUpdateFlag.

virtual void itk::ProcessObject::PrepareOutputs ( ) [virtual]

An opportunity to deallocate a ProcessObject's bulk data storage. Some filters may wish to reuse existing bulk data storage to avoid unnecessary deallocation/allocation sequences. The default implementation calls Initialize() on each output. DataObject::Initialize() frees its bulk data by default.
Reimplemented in itk::WatershedImageFilter< TInputImage >.

virtual void itk::ProcessObject::PropagateRequestedRegion ( DataObject * output ) [virtual]

Send the requested region information back up the pipeline (to the filters that preceed this one).
Reimplemented in itk::VTKImageImport< TOutputImage >.

virtual void itk::ProcessObject::PropagateResetPipeline ( ) [protected, virtual]

Propagate a call to ResetPipeline() up the pipeline. Called only from DataObject.

virtual void itk::ProcessObject::ReleaseDataBeforeUpdateFlagOff ( ) [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released/reallocated during an Update(). In limited memory scenarios, a user may want to force the elements of a pipeline to release any bulk data that is going to be regenerated anyway during an Update() in order to control peak memory allocation. Note that this flag is different from the ReleaseDataFlag. ReleaseDataFlag manages the deallocation of a ProcessObject's bulk output data once that data has been consumed by a downstream ProcessObject. The ReleaseDataBeforeUpdateFlag manages the deallocation/reallocation of bulk data during a pipeline update to control peak memory utilization. Default value is on.

virtual void itk::ProcessObject::ReleaseDataBeforeUpdateFlagOn ( ) [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released/reallocated during an Update(). In limited memory scenarios, a user may want to force the elements of a pipeline to release any bulk data that is going to be regenerated anyway during an Update() in order to control peak memory allocation. Note that this flag is different from the ReleaseDataFlag. ReleaseDataFlag manages the deallocation of a ProcessObject's bulk output data once that data has been consumed by a downstream ProcessObject. The ReleaseDataBeforeUpdateFlag manages the deallocation/reallocation of bulk data during a pipeline update to control peak memory utilization. Default value is on.

void itk::ProcessObject::ReleaseDataFlagOff ( ) [inline]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released after being used by a downstream ProcessObject. Default value is off. Another options for controlling memory utilization is the ReleaseDataBeforeUpdateFlag.
Definition at line 254 of file itkProcessObject.h.

void itk::ProcessObject::ReleaseDataFlagOn ( ) [inline]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released after being used by a downstream ProcessObject. Default value is off. Another options for controlling memory utilization is the ReleaseDataBeforeUpdateFlag.
Definition at line 253 of file itkProcessObject.h.

virtual void itk::ProcessObject::RemoveInput ( DataObject * input ) [protected, virtual]

Protected methods for setting inputs. Subclasses make use of them for setting input.

virtual void itk::ProcessObject::RemoveOutput ( DataObject * output ) [protected, virtual]

Protected methods for setting outputs. Subclasses make use of them for getting output.

virtual void itk::ProcessObject::ResetPipeline ( ) [virtual]

Reset the pipeline. If an exception is thrown during an Update(), the pipeline may be in an inconsistent state. This method clears the internal state of the pipeline so Update() can be called.

virtual void itk::ProcessObject::RestoreInputReleaseDataFlags ( ) [protected, virtual]

Restore the cached input ReleaseDataFlags.

virtual void itk::ProcessObject::SetAbortGenerateData ( bool _arg ) [virtual]

Set the AbortGenerateData flag for the process object. Process objects may handle premature termination of execution in different ways.

virtual void itk::ProcessObject::SetNthInput ( unsigned int num,

DataObject * input

) [protected, virtual]

Protected methods for setting inputs. Subclasses make use of them for setting input.

virtual void itk::ProcessObject::SetNthOutput ( unsigned int num,

DataObject * output

) [protected, virtual]

Protected methods for setting outputs. Subclasses make use of them for getting output.

void itk::ProcessObject::SetNumberOfInputs ( unsigned int num ) [protected]

Called to allocate the input array. Copies old inputs.

void itk::ProcessObject::SetNumberOfOutputs ( unsigned int num ) [protected]

Called to allocate the output array. Copies old outputs.

virtual void itk::ProcessObject::SetNumberOfRequiredInputs ( unsigned int _arg ) [protected, virtual]

Protected methods for setting inputs. Subclasses make use of them for setting input.

virtual void itk::ProcessObject::SetNumberOfRequiredOutputs ( unsigned int _arg ) [protected, virtual]

Protected methods for setting outputs. Subclasses make use of them for getting output.

virtual void itk::ProcessObject::SetNumberOfThreads ( int _arg ) [virtual]

Get/Set the number of threads to create when executing.

virtual void itk::ProcessObject::SetProgress ( float _arg ) [virtual]

Set the execution progress of a process object. The progress is a floating number in [0,1] with 0 meaning no progress and 1 meaning the filter has completed execution. The ProgressEvent is NOT invoked.

virtual void itk::ProcessObject::SetReleaseDataBeforeUpdateFlag ( bool _arg ) [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released/reallocated during an Update(). In limited memory scenarios, a user may want to force the elements of a pipeline to release any bulk data that is going to be regenerated anyway during an Update() in order to control peak memory allocation. Note that this flag is different from the ReleaseDataFlag. ReleaseDataFlag manages the deallocation of a ProcessObject's bulk output data once that data has been consumed by a downstream ProcessObject. The ReleaseDataBeforeUpdateFlag manages the deallocation/reallocation of bulk data during a pipeline update to control peak memory utilization. Default value is on.

virtual void itk::ProcessObject::SetReleaseDataFlag ( bool flag ) [virtual]

Turn on/off the flags to control whether the bulk data belonging to the outputs of this ProcessObject are released after being used by a downstream ProcessObject. Default value is off. Another options for controlling memory utilization is the ReleaseDataBeforeUpdateFlag.

virtual void itk::ProcessObject::Update ( ) [virtual]

Bring this filter up-to-date. Update() checks modified times against last execution times, and re-executes objects if necessary. A side effect of this method is that the whole pipeline may execute in order to bring this filter up-to-date. This method updates the currently prescribed requested region. If no requested region has been set on the output, then the requested region will be set to the largest possible region. Once the requested region is set, Update() will make sure the specified requested region is up-to-date. This is a confusing side effect to users who are just calling Update() on a filter. A first call to Update() will cause the largest possible region to be updated. A second call to Update() will update that same region. If a modification to the upstream pipeline cause a filter to have a different largest possible region, this second call to Update() will not cause the output requested region to be reset to the new largest possible region. Instead, the output requested region will be the same as the last time Update() was called. To have a filter always to produce its largest possible region, users should call UpdateLargestPossibleRegion() instead.
Reimplemented in itk::CoreAtomImageToUnaryCorrespondenceMatrixProcess< TSourceImage >, itk::MedialNodePairCorrespondenceProcess< TSourceImage >, itk::MedialNodeTripletCorrespondenceProcess< TSourceImage >, itk::CoreAtomImageToDistanceMatrixProcess< TSourceImage >, itk::ImageFileWriter< TInputImage >, and itk::ImageSeriesWriter< TInputImage, TOutputImage >.

virtual void itk::ProcessObject::UpdateLargestPossibleRegion ( ) [virtual]

Like Update(), but sets the output requested region to the largest possible region for the output. This is the method users should call if they want the entire dataset to be processed. If a user wants to update the same output region as a previous call to Update() or a previous call to UpdateLargestPossibleRegion(), then they should call the method Update().

virtual void itk::ProcessObject::UpdateOutputData ( DataObject * output ) [virtual]

Actually generate new output
Reimplemented in itk::StreamingImageFilter< TInputImage, TOutputImage >.

virtual void itk::ProcessObject::UpdateOutputInformation ( ) [virtual]

Update the information decribing the output data. This method transverses up the pipeline gathering modified time information. On the way back down the pipeline, this method calls GenerateOutputInformation() to set any necessary information about the output data objects. For instance, a filter that shrinks an image will need to provide an implementation for GenerateOutputInformation() that changes the spacing of the pixels. Such filters should call their superclass' implementation of GenerateOutputInformation prior to changing the information values they need (i.e. GenerateOutputInformation() should call Superclass::GenerateOutputInformation() prior to changing the information.
Reimplemented in itk::watershed::Segmenter< TInputImage >, itk::VTKImageImport< TOutputImage >, and itk::watershed::Segmenter< InputImageType >.

void itk::ProcessObject::UpdateProgress ( float amount )

Update the progress of the process object.
Sets the Progress ivar to amount and invokes any observers for the ProgressEvent. The parameter amount should be in [0,1] and is the cumulative (not incremental) progress.

Friends And Related Function Documentation

friend class DataObject [friend]

Friends of ProcessObject
Definition at line 446 of file itkProcessObject.h.

Member Data Documentation

TimeStamp itk::ProcessObject::m_OutputInformationMTime [protected]

Time when GenerateOutputInformation was last called.
Definition at line 416 of file itkProcessObject.h.

bool itk::ProcessObject::m_Updating [protected]

This flag indicates when the pipeline is executing. It prevents infinite recursion when pipelines have loops.
Definition at line 413 of file itkProcessObject.h.

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

itkProcessObject.h

Generated at Thu May 25 02:08:24 2006 for ITK by

itk::ProcessObject Class Reference [ITK System Objects, Data Processing Objects]

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes