goDWT< T > Class Template Reference

#include <godwt.h>

Public Member Functions
void	haar (goSignal3D< T > &signal, goSignal3D< godwt_t > &targetSignal)
int	haar (goSignal3D< T > &signal, goSignal3D< godwt_t > &targetSignal, int stage)
void	haar (goSubSignal3D< T > &signal)
int	haar (goSignal3D< T > &signal, int stage)
void	unHaar (goSignal3D< godwt_t > &haarSignal, goSignal3D< T > &targetSignal)
void	unHaar (goSubSignal3D< T > &haarSignal)
int	unHaar (goSignal3D< T > &signal, int stage)

Detailed Description

template<class T>
class goDWT< T >

DWT class for 3d signals (separable, so it's for all signals dim < 4)

Author

Christian Gosch

Todo:

src/signal/godwt.cc aufraeumen

Examples

dwt.cc.

Member Function Documentation

haar() [1/3]

template<class T >

void goDWT< T >::haar	(	goSignal3D< T > &	signal,
		goSignal3D< godwt_t > &	targetSignal
	)

Non-specialised Haar transform. Uses goDouble values for the transform,

Parameters

signal	The signal to be transformed.
targetSignal	The signal where the dwt is to be stored in. targetSignal has to be made using goSignal3D::make() in the correct size.

See also

goDWT::unHaar()

haar() [2/3]

template<class T >

int goDWT< T >::haar	(	goSignal3D< T > &	signal,
		int	stage
	)

In place version of the Haar transform with stage stages. This is implemented for the types

• goInt8
• goUInt8
• goInt16
• goUInt16
• goInt32
• goUInt32
• goFloat
• goDouble The integer type transform is realised not as a "real" Haar transform, but as a modified S-Transform as described in B. Aiazzi et al., "Lossless compression of medical images based on an enhanced generalized multidimensional S-Transform", SPIE Vol. 2825, pp. 868–878.

  Parameters

  signal	Contains the signal to be transformed and after execution, contains the transform of the signal.
  stage	Stage up to which the transform shall be executed.

haar() [3/3]

template<class T>

void goDWT< T >::haar

(

goSubSignal3D< T > &

signal

)

In place version of the Haar transform. There are special implementations for goFloat, goDouble, ...
Note that this can only be used for types allowing for the type of the transform to be the same as the signal itself. Currently, this is possible for goFloat and goDouble.

Parameters

signal

Contains the signal to be transformed and after execution, contains the transform of the signal.

unHaar() [1/3]

template<class T >

void goDWT< T >::unHaar	(	goSignal3D< godwt_t > &	haarSignal,
		goSignal3D< T > &	targetSignal
	)

Non-specialised Haar reverse transform. Pretty slow due to a copy operation at the end.

Todo:

Remove the copy operation at the end by specialising a macro to the last filter operation. Using godwt_t types for all data is better anyway as long as it is scalar data. Use a different solution for multimodal data.

Parameters

haarSignal	Signal containing the Haar transform as produced by haar()
targetSignal	Signal where the reverse transform is to be stored in. Again, this has to be created before calling this member.

See also

goDWT::haar()

unHaar() [2/3]

template<class T >

int goDWT< T >::unHaar	(	goSignal3D< T > &	signal,
		int	stage
	)

See also

haar(goSignal3D, int);

unHaar() [3/3]

template<class T>

void goDWT< T >::unHaar

(

goSubSignal3D< T > &

haarSignal

)

In place version of the reverse Haar transform. There are special implementations for goFloat, ...

---

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

• godwt.h
• godwt.cc