lti::convHelper2D< T, A > Class Template Reference

pre-convolution algorithm for a matrix * kernel2D The type T is the type of the elements of the vector an kernel1D The class A determines how the values should be accumulated. More...

#include <ltiConvolutionHelper.h>

Collaboration diagram for lti::convHelper2D< T, A >:

[legend]

List of all members.

Public Member Functions
	convHelper2D ()
	~convHelper2D ()
bool	setKernel (const mathObject *kern)
bool	isKernelSeparable () const
void	apply (const matrix< T > &src, matrix< T > &dest, const modifier::parameters &param)
void	applySep (const matrix< T > &src, matrix< T > &dest, const modifier::parameters &param)
void	applyMask (const matrix< T > &src, matrix< T > &dest, const modifier::parameters &param)
void	setMask (const channel8 &mask)
const channel8 *	getMask () const
A &	getAccumulator ()
const A &	getAccumulator () const

---

Detailed Description

template<class T, class A>
class lti::convHelper2D< T, A >

pre-convolution algorithm for a matrix * kernel2D The type T is the type of the elements of the vector an kernel1D The class A determines how the values should be accumulated.

The class A must have following methods:

• accumulate(T kernelElement,T source) accumulate the result of applying some function to one element of the kernel and one element of the source channel.
• accumulateSym(T kernelElement,T source1, T source2) accumulate the result of applying some function to one element of a symmetric kernel and the corresponding elements of the source channel.
• accumulateASym(T kernelElement,T source1, T source2) accumulate the result of applying some function to one element of a anti-symmetric kernel (e.g. a gradient) and the corresponding elements of the source channel.
• accumulateZero(T source) accumulate the result of applying some function with kernel value zero to the source channel. For efficient implementation of accumulateASym() where middle kernel value is zero by definition. For filters this is an empty function, however, not for others.
• getResult() returns the accumulated value
• reset() reset the accumulated value
• setNorm() set the normalization constant used by the getResult() method

Examples of the accumulator class can be found in: lti::dilation::accumulatorGray, lti::dilation::accumulatorBin, lti::erosion::accumulatorGray, lti::erosion::accumulatorBin, lti::convolution::accumulator

The implementation of these methods allows the efficient use of this class for the implementation of convolution, dilation and erosion.

Please note that this class is intended to be used just as a helper class. The correctness of the parameters and input data must be ensured by the "helped" classes before these methods are called.

---

Constructor & Destructor Documentation

template<class T , class A >

lti::convHelper2D< T, A >::convHelper2D

(

)

default constructor

template<class T , class A >

lti::convHelper2D< T, A >::~convHelper2D

(

)

destructor

---

Member Function Documentation

template<class T , class A >

void lti::convHelper2D< T, A >::apply	(	const matrix< T > &	src,
		matrix< T > &	dest,
		const modifier::parameters &	param
	)

convHelpers the matrix src with the filter kernel and leaves the result in dest.

Consistency checks should be done by the apply methods!

Parameters:

	src	matrix to be filtered
	dest	result
	param	parameters with the boundaryType to be used

template<class T , class A >

void lti::convHelper2D< T, A >::applyMask	(	const matrix< T > &	src,
		matrix< T > &	dest,
		const modifier::parameters &	param
	)

convolves the matrix src with the object depicted in the filter kernel and described by the mask.

Leaves the result in dest.

Consistency checks should be done by the apply methods!

Parameters:

	src	matrix to be filtered
	dest	result
	param	parameters that determine e.g. border behaviour

template<class T , class A >

void lti::convHelper2D< T, A >::applySep	(	const matrix< T > &	src,
		matrix< T > &	dest,
		const modifier::parameters &	param
	)

convHelpers the matrix src with the separable filter kernel and leaves the result in dest.

Consistency checks should be done by the apply methods!

Parameters:

	src	matrix to be filtered
	dest	result
	param	parameters with the boundaryType to be used

template<class T , class A >

const A& lti::convHelper2D< T, A >::getAccumulator

(

)

const

returns a reference to the accumulator object being used

template<class T , class A >

A& lti::convHelper2D< T, A >::getAccumulator

(

)

returns a reference to the accumulator object being used

template<class T , class A >

const channel8* lti::convHelper2D< T, A >::getMask

(

)

const

Returns a pointer to the current mask.

template<class T , class A >

bool lti::convHelper2D< T, A >::isKernelSeparable

(

)

const [inline]

get kernel type

template<class T , class A >

bool lti::convHelper2D< T, A >::setKernel

(

const mathObject *

kern

)

try to get a Kernel from the parameters.

If it is not possible to get it, return false; This will not copy the kernel, it just will keep a reference to it!

Parameters:

kern

the mathObject with the kernel.

Returns:

true if kernel could be set and false otherwise

template<class T , class A >

void lti::convHelper2D< T, A >::setMask

(

const channel8 &

mask

)

Set the mask to be used with the kernel.

---

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

• ltiConvolutionHelper.h