Instances define a lifting step in a wavelet lifting scheme. More...

#include <Wavelet.h>

Inheritance diagram for bwave::Wavelet::LiftingStep:

Public Types
enum	Type { PRIMAL, DUAL, SSCALE, DSCALE }
	Constants to model lifting step types.
Public Member Functions
Type	type () const
	Get type of lifting.
int	origin () const
	Get the origin of the lifting step filter.
int	size () const
	Get the lifting step filter size.
double	coeff (int i) const
	Get a filter coefficient.
double	divisor () const
	Get the divisor.
template<class tp_Type >
void	apply (blitz::Array< tp_Type, 1 > &s, blitz::Array< tp_Type, 1 > &d, bool inverse, ExtensionMode be=CONSTANT_EXT) const
	Apply the lifting step on some data.
	LiftingStep ()
	Default constructor.
	LiftingStep (Type t, int o, double d, double c0)
	Construct a LiftingStep from it's specification.
	LiftingStep (Type t, int o, double d, double c0, double c1)
	LiftingStep (Type t, int o, double d, double c0, double c1, double c2)
	LiftingStep (Type t, int o, double d, double c0, double c1, double c2, double c3)
	LiftingStep (Type t, int o, double d, double c0, double c1, double c2, double c3, double c4)
	LiftingStep (Type t, int o, double d, double c0, double c1, double c2, double c3, double c4, double c5)
bool	operator== (const LiftingStep &s) const
	Compare two lifting steps.
virtual	~LiftingStep ()
	Destructor.
Protected Member Functions
template<class tp_Type , int tp_size>
void	applyFS (blitz::Array< tp_Type, 1 > &s, blitz::Array< tp_Type, 1 > &d, bool inverse, ExtensionMode be) const
Protected Attributes
Type	m_type
int	m_origin
std::vector< double >	m_coeffs
double	m_divisor

Detailed Description

Instances define a lifting step in a wavelet lifting scheme.

A lifting step consists of a filter and a divisor. When applied, the lifting step filters one of two given data channels (s and d), divides the filtered values by the divisor and then adds/subtracts them to/from the data in the other channel. The divisor is necessary for integer lifting, use only integer filter coefficients and integer divisor for integer compatibility. You can use such an integer compatible lifting for both integer and floating point data-types. Divisors which are powers of two will result in much faster computation.

In addition to classical primal/dual lifting steps you can specify scaling steps to scale the d- or s-channel. However, scaling steps lead to information loss in at least one transformation direction when operating on integer data.

Constructor & Destructor Documentation

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0
	)

Construct a LiftingStep from it's specification.

Parameters:

t	PRIMAL for a primal, DUAL for a dual lifting step, SSCALE, DSCALE for a scaling step.
o	Filter origin.
d	Divisor.
c0	First filter coefficient, use `c1`, `c2`, etc. for more.

Use only this constructor (i.e. use only one filter coefficient) to specify a scaling step.

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0,
		double	c1
	)

See also:: LiftingStep (Type t, int o, double d, double c0)

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0,
		double	c1,
		double	c2
	)

See also:: LiftingStep (Type t, int o, double d, double c0)

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0,
		double	c1,
		double	c2,
		double	c3
	)

See also:: LiftingStep (Type t, int o, double d, double c0)

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0,
		double	c1,
		double	c2,
		double	c3,
		double	c4
	)

See also:: LiftingStep (Type t, int o, double d, double c0)

bwave::Wavelet::LiftingStep::LiftingStep	(	Type	t,
		int	o,
		double	d,
		double	c0,
		double	c1,
		double	c2,
		double	c3,
		double	c4,
		double	c5
	)

See also:: LiftingStep (Type t, int o, double d, double c0)

Member Function Documentation

template<class tp_Type >

void bwave::Wavelet::LiftingStep::apply	(	blitz::Array< tp_Type, 1 > &	s,
		blitz::Array< tp_Type, 1 > &	d,
		bool	inverse,
		ExtensionMode	be = `CONSTANT_EXT`
	)		const `[inline]`

Apply the lifting step on some data.

Parameters:

s	"Low-frequency-" or "Phi-" data channel.
d	"High-frequency-" or "Psi-" data channel.
inverse	Do inverse lifting.
be	Use this extension mode for boundary handling.

Referenced by bwave::Wavelet::invLift(), and bwave::Wavelet::lift().

template<class tp_Type , int tp_size>

void bwave::Wavelet::LiftingStep::applyFS	(	blitz::Array< tp_Type, 1 > &	s,
		blitz::Array< tp_Type, 1 > &	d,
		bool	inverse,
		ExtensionMode	be
	)		const `[inline, protected]`

! Add speed-up for divisor()==2^n

double bwave::Wavelet::LiftingStep::coeff ( int i ) const [inline]

Get a filter coefficient.

Parameters:

i	index of the coefficient (i>=0)

Returns:: The i-th filter coefficient.

Referenced by operator==().

double bwave::Wavelet::LiftingStep::divisor ( ) const [inline]

Get the divisor.

Returns:: The lifting step divisor.

Referenced by operator==().

bool bwave::Wavelet::LiftingStep::operator== ( const LiftingStep & s ) const

Compare two lifting steps.

Parameters:

s	target to compare with

Returns:: true if step types, filters and divisors are equal, false if not.

References coeff(), divisor(), bwave::Wavelet::equals(), origin(), size(), and type().

int bwave::Wavelet::LiftingStep::origin ( ) const [inline]

Get the origin of the lifting step filter.

Returns:: The relative position of the first filter coefficient for the convolution.

Referenced by operator==().

int bwave::Wavelet::LiftingStep::size ( ) const [inline]

Get the lifting step filter size.

Returns:: The number of filter coefficients.

Referenced by operator==().

Type bwave::Wavelet::LiftingStep::type ( ) const [inline]

Get type of lifting.

Returns:: PRIMAL for a primal, DUAL for a dual lifting step, SSCALE, DSCALE for a scaling step.

Referenced by operator==().

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

/home/pclz-02/oschulz/Data/Science/Projects/Gerda/Software/blitzwave/src/Wavelet.h
/home/pclz-02/oschulz/Data/Science/Projects/Gerda/Software/blitzwave/src/Wavelet.cxx

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation