CTL::OCL::ApodizationFilter Class Reference

The ApodizationFilter class is an OpenCL implementation of several apodization functions. More...

#include <oclprojectionfilters.h>

Inheritance diagram for CTL::OCL::ApodizationFilter:

Collaboration diagram for CTL::OCL::ApodizationFilter:

Public Types
enum	{ Type = 3103 }
enum	FilterType { RamLak, SheppLogan, Cosine, Hann }
Public Types inherited from CTL::OCL::GenericOCLProjectionFilter
enum	{ Type = 3100 }
Public Types inherited from CTL::AbstractProjectionFilter
enum	{ Type = 3000 }
Public Types inherited from CTL::SerializationInterface
enum	{ Type = -1, UserType = 65536 }

Public Member Functions
int	type () const override
void	filter (ProjectionData &projections) override
	Applies the apodization filter operation to projections.
	ApodizationFilter (FilterType filterType, float frequencyScaling=1.0f)
	Constructs an ApodizationFilter object. More...
QVariant	parameter () const override
	Returns the parameters of this instance as a QVariant. More...
void	setParameter (const QVariant &parameter) override
	Sets the parameters of this instance based on the passed QVariant parameter. More...
Public Member Functions inherited from CTL::OCL::GenericOCLProjectionFilter
	GenericOCLProjectionFilter (const std::string &clFileName, const std::vector< float > &arguments={})
	Creates a GenericOCLProjectionFilter instance that is capable of executing the OpenCL kernel (filter) of the file clFileName. More...
void	setAdditionalKernelArg (float argument)
	Convenience method for simpler setting of a single additional argument. More...
void	setAdditionalKernelArgs (const std::vector< float > &arguments)
	Sets the additional kernel arguments to arguments. More...
Public Member Functions inherited from CTL::AbstractProjectionFilter
void	fromVariant (const QVariant &variant) override
QVariant	toVariant () const override
Public Member Functions inherited from CTL::SerializationInterface
virtual	~SerializationInterface ()=default

Private Member Functions
	ApodizationFilter ()
	Default ctor; private. Used for purpose of deserialization only.
std::vector< float >	generateCoefficients (unsigned filterSize) const
	Generates the coefficients for a filter of size filterSize.

Static Private Member Functions
static QMetaEnum	metaEnum ()
	QMetaEnum object used to decode/encode subset order enum values.

Private Attributes
FilterType	_filterType
float	_frequencyScaling = 1.0f
std::vector< float >	_filter

Friends
template<class >
struct	SerializationHelper::RegisterWithSerializationHelper

Additional Inherited Members
Protected Member Functions inherited from CTL::OCL::GenericOCLProjectionFilter
	GenericOCLProjectionFilter ()
	Creates a GenericOCLProjectionFilter without any OpenCL kernel associated to it. More...
	GenericOCLProjectionFilter (cl::Kernel *kernel, const std::vector< float > &arguments={})
	Creates a GenericOCLProjectionFilter that executes the OpenCL kernel kernel. More...
virtual cl::NDRange	globalWorksize (const ProjectionData &projections) const
	Returns the global worksize for the kernel call. More...
virtual cl::NDRange	localWorksize (const ProjectionData &projections) const
	Returns the local worksize for the kernel call. More...
virtual bool	autoCombine () const
	Returns true if projections will be automatically combined before being processed by the kernel. More...
Protected Member Functions inherited from CTL::AbstractProjectionFilter
	AbstractProjectionFilter (const AbstractProjectionFilter &)=default
	AbstractProjectionFilter (AbstractProjectionFilter &&)=default
AbstractProjectionFilter &	operator= (const AbstractProjectionFilter &)=default
AbstractProjectionFilter &	operator= (AbstractProjectionFilter &&)=default
Protected Member Functions inherited from CTL::SerializationInterface
	SerializationInterface ()=default
	SerializationInterface (const SerializationInterface &)=default
	SerializationInterface (SerializationInterface &&)=default
SerializationInterface &	operator= (const SerializationInterface &)=default
SerializationInterface &	operator= (SerializationInterface &&)=default
Protected Attributes inherited from CTL::OCL::GenericOCLProjectionFilter
std::string	_clFileName
std::vector< float >	_additionalArgs
cl::Kernel *	_kernel
cl::CommandQueue	_queue

Detailed Description

The ApodizationFilter class is an OpenCL implementation of several apodization functions.

For direct reconstructions of volumes from X-ray projections, e.g. using FDKReconstructor, the Ram-Lak filter is applied on the projections in frequency domain. Due to the discrete nature of detector pixels, the Ram-Lak filter is band-limited and sampled at discrete positions, which leads to spectral leakage. Moreover, in case of a low signal-to-noise ratio of the projections, it is sensible to attenuate high frequency content to reduce noise in the reconstructions. Apodization functions, also known as window functions, are used to handle these two problems. They are (usually) symmetric functions that are zero-valued outside a certain interval. For CT reconstruction, they are multiplied with the Ram-Lak filter in frequency domain.

Member Enumeration Documentation

FilterType

enum CTL::OCL::ApodizationFilter::FilterType

Enumerator
RamLak	No apodization. This is the only apodization type that does not need to calculate the discrete Fourier transform and its inverse if \(s=1\).
SheppLogan	Multiplies the Ram-Lak filter by the sinc function. \[ \mathrm{apod}(\omega) = \frac{\sin(s^{-1}\omega)}{s^{-1}\omega} \] L. A. Shepp and B. F. Logan, "The Fourier reconstruction of a head section," in IEEE Transactions on Nuclear Science, vol. 21, no. 3, pp. 21-43, June 1974, doi: 10.1109/TNS.1974.6499235.
Cosine	Multiplies the Ram-Lak filter by the cosine function. \[ \mathrm{apod}(\omega) = \cos(0.5s^{-1}\omega) \] F. J. Harris, "On the use of windows for harmonic analysis with the discrete Fourier transform," in Proceedings of the IEEE, vol. 66, no. 1, pp. 51-83, Jan. 1978, doi: 10.1109/PROC.1978.10837.
Hann	Multiplies the Ram-Lak filter by the Hann window. \[ \mathrm{apod}(\omega) = \frac{1}{2}(1 + \cos(s^{-1}\omega)) \] R. B. Blackman and J. W. Tukey, "The measurement of power spectra from the point of view of communications engineering — Part I," in The Bell System Technical Journal, vol. 37, no. 1, pp. 185-282, Jan. 1958, doi: 10.1002/j.1538-7305.1958.tb03874.x.

Constructor & Destructor Documentation

ApodizationFilter()

CTL::OCL::ApodizationFilter::ApodizationFilter ( FilterType  filterType, float  frequencyScaling = 1.0f  )

explicit

Constructs an ApodizationFilter object.

Parameters

filterType	Type of the apodization filter
frequencyScaling	Scaling of the apodization function along frequency axis. Can take values in \((0, 1]\). Frequencies above/below the cutoff frequency calculated from this scaling factor are set to zero.

Member Function Documentation

parameter()

QVariant CTL::OCL::ApodizationFilter::parameter ( ) const

overridevirtual

Returns the parameters of this instance as a QVariant.

Returns a QVariantMap including the following (key, value)-pairs:

• information from the base class (GenericOCLProjectionFilter),
• ("frequency scaling", [float] scaling of the apodization function),
• ("filter type", [string] name of the filter type).

Reimplemented from CTL::OCL::GenericOCLProjectionFilter.

setParameter()

void CTL::OCL::ApodizationFilter::setParameter ( const QVariant &  parameter )

overridevirtual

Sets the parameters of this instance based on the passed QVariant parameter.

The passed parameter must be a QVariantMap containing one or more of the following (key, value)-pairs:

• information from the base class (GenericOCLProjectionFilter),
• ("frequency scaling", [float] scaling of the apodization function),
• ("filter type", [string] name of the filter type).

Note that changing parameters of the base class this way (esp. changing the OpenCL kernel file) leads to undefined behavior.

Reimplemented from CTL::OCL::GenericOCLProjectionFilter.

type()

int CTL::OCL::ApodizationFilter::type ( ) const

inlineoverridevirtual

Returns the type-id of the serializable object. Used in deserialization to determine the proper object type.

Add derived classes to the enumeration using the CTL_TYPE_ID macro.

Reimplemented from CTL::OCL::GenericOCLProjectionFilter.

---

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

• modules/src/processing/oclprojectionfilters.h
• modules/src/processing/oclprojectionfilters.cpp