lti::sepKernel< T > Class Template Reference

Separable Kernel. More...

#include <ltiLinearKernels.h>

Inheritance diagram for lti::sepKernel< T >:

[legend]

Collaboration diagram for lti::sepKernel< T >:

[legend]

List of all members.

Public Member Functions
	sepKernel ()
	sepKernel (const sepKernel &other)
	sepKernel (const int &from, const int &to, const T &iniValue=T())
	sepKernel (const kernel1D< T > &subkernel)
virtual	~sepKernel ()
sepKernel &	copy (const sepKernel &other)
const char *	getTypeName () const
virtual mathObject *	clone () const
template<class U >
sepKernel &	castFrom (const sepKernel< U > &other)
bool	separate (const kernel2D< T > &k, const double &maxDev=0.01)
int	getNumberOfPairs () const
void	setNumberOfPairs (const int &numPairs)
kernel1D< T > &	getRowFilter (const int &i)
kernel1D< T > &	getColFilter (const int &i)
const kernel1D< T > &	getRowFilter (const int &i) const
const kernel1D< T > &	getColFilter (const int &i) const
void	denormalize ()
sepKernel< T > &	multiply (const T &value)
void	setNorm (const T &newNorm)
T	sumOfElements () const
sepKernel< T > &	mirror (const sepKernel< T > &other)
sepKernel< T > &	mirror ()
virtual bool	write (ioHandler &handler, const bool complete=true) const
virtual bool	read (ioHandler &handler, const bool complete=true)
Apply Methods
sepKernel< T > &	apply (T(*function)(T))
sepKernel< T > &	apply (T(*function)(const T &))
sepKernel< T > &	apply (const sepKernel< T > &other, T(*function)(T))
sepKernel< T > &	apply (const sepKernel< T > &other, T(*function)(const T &))
Protected Attributes
std::vector< kernel1D< T > >	rowKernels
std::vector< kernel1D< T > >	colKernels

---

Detailed Description

template<class T>
class lti::sepKernel< T >

Separable Kernel.

A separable kernel is a vector of one dimensional kernels. If a two dimensional kernel can be separated, the convolution can be applied in a very efficient way.

A filter kernel K is called separable "in one pair", if the matrix representation of K can be produced as an outer product of two one-dimensional kernels Kx and Ky.

The template type of this class should coincide with the template class of the matrix or channel to be convolved with. For example, if you want to convolve a sepKernel with a lti::channel, you will need a sepKernel<channel::value_type> or sepKernel<float>.

If you instantiate a sepKernel of a fixed point type, like sepKernel<int> or sepKernel<ubyte>, you also need to consider the "norm" of the kernel (see lti::sepKernel<T>::getNorm() and lti::sepKernel<T>::setNorm(const T&)). This "norm" allows the representation of numbers less than 1.0. You can see this norm as the value to be consider as 1.0, when operating with the kernel. For example, if you have a sepKernel<ubyte> with the values [64,128,64] and norm=255, the the interpreted values during convolution will be [0.25,0.5,0.25]. With floating-point types, the norm will be always assumed to be 1.0. For other types the default norms are the following: (see lti::typeInfo)

• For int: 65536
• For ubyte: 255
• Otherwise: 1.0

See also:

convolution

---

Constructor & Destructor Documentation

template<class T>

lti::sepKernel< T >::sepKernel

(

)

default constructor

template<class T>

lti::sepKernel< T >::sepKernel

(

const sepKernel< T > &

other

)

copy constructor

Parameters:

other

the kernel to be copied.

template<class T>

lti::sepKernel< T >::sepKernel	(	const int &	from,
		const int &	to,
		const T &	iniValue = T()
	)

construct a separable kernel with one filter pair, all elements of the subfilters initialized with the given value.

Parameters:

	from	first index of the one dimensional filter kernel
	to	last index of the one dimensional filter kernel
	iniValue	initial value for the kernel elements

template<class T>

lti::sepKernel< T >::sepKernel

(

const kernel1D< T > &

subkernel

)

construct a symmetrical separable kernel

The resulting separable kernel will have just one filter pair, where the row and column filters are identical.

Parameters:

subkernel

the one-dimensional kernel to be used as row and column filter.

template<class T>

virtual lti::sepKernel< T >::~sepKernel

(

)

[virtual]

destructor

---

Member Function Documentation

template<class T>

sepKernel<T>& lti::sepKernel< T >::apply	(	const sepKernel< T > &	other,
		T(*)(const T &)	function
	)

applies a C-function to each element the other kernel and leaves the result here.

Parameters:

	other	the kernel with the source data
	function	a pointer to a C-function

Returns:

a reference to the actual kernel

template<class T>

sepKernel<T>& lti::sepKernel< T >::apply	(	const sepKernel< T > &	other,
		T(*)(T)	function
	)

applies a C-function to each element of the other kernel and leaves the result here.

Parameters:

	other	the source kernel
	function	a pointer to a C-function

Returns:

a reference to the actual kernel

template<class T>

sepKernel<T>& lti::sepKernel< T >::apply

(

T(*)(const T &)

function

)

applies a C-function to each element of the kernel.

Parameters:

function

a pointer to a C-function

Returns:

a reference to the actual kernel

template<class T>

sepKernel<T>& lti::sepKernel< T >::apply

(

T(*)(T)

function

)

applies a C-function to each element of the kernel.

In the following example, kernel kernel is initialized with 4.0. After applying sqrt(), all elements of kernel are 2.0.

 sepKernel<float> kern(-2,2,4.0);
 kern.apply(sqrt);

Parameters:

function

a pointer to a C-function

Returns:

a reference to the actual kernel

template<class T>

template<class U >

sepKernel& lti::sepKernel< T >::castFrom

(

const sepKernel< U > &

other

)

[inline]

copy from kernel of another type

Parameters:

other

a separable kernel of another type

Returns:

a reference to this instance

template<class T>

virtual mathObject* lti::sepKernel< T >::clone

(

)

const [inline, virtual]

clone member

Returns:

a pointer to a copy of this instance

Implements lti::mathObject.

template<class T>

sepKernel& lti::sepKernel< T >::copy

(

const sepKernel< T > &

other

)

copy member

Parameters:

other

the kernel to be copied.

Returns:

a reference to this instance

Reimplemented from lti::ioObject.

template<class T>

void lti::sepKernel< T >::denormalize

(

)

denormalize divide all elements by norm and set the norm to 1!

template<class T>

const kernel1D<T>& lti::sepKernel< T >::getColFilter

(

const int &

i

)

const [inline]

return an unmodifiable column kernel

Parameters:

i

the index of the column filter. This value must be between 0 and getNumberOfPairs()

template<class T>

kernel1D<T>& lti::sepKernel< T >::getColFilter

(

const int &

i

)

[inline]

return a column-kernel

Parameters:

i

the index of the column filter. This value must be between 0 and getNumberOfPairs()

Referenced by lti::sepKernel< float >::castFrom().

template<class T>

int lti::sepKernel< T >::getNumberOfPairs

(

)

const

number of filter pairs

Referenced by lti::sepKernel< float >::castFrom().

template<class T>

const kernel1D<T>& lti::sepKernel< T >::getRowFilter

(

const int &

i

)

const [inline]

return an unmodifiable row kernel

Parameters:

i

the index of the row filter. This value must be between 0 and getNumberOfPairs()

template<class T>

kernel1D<T>& lti::sepKernel< T >::getRowFilter

(

const int &

i

)

[inline]

return a row-kernel

Parameters:

i

the index of the row filter. This value must be between 0 and getNumberOfPairs()

Referenced by lti::sepKernel< float >::castFrom().

template<class T>

const char* lti::sepKernel< T >::getTypeName

(

void

)

const [inline, virtual]

returns name of this type ("sepKernel")

Reimplemented from lti::mathObject.

template<class T>

sepKernel<T>& lti::sepKernel< T >::mirror

(

)

Mirror this kernel, i.e.

at(y,x) = at(-y,-x);

Returns:

a reference to this instance

template<class T>

sepKernel<T>& lti::sepKernel< T >::mirror

(

const sepKernel< T > &

other

)

Mirror the other kernel and leave the result here, i.e.

at(y,x) = other.at(-y,-x);

Parameters:

other

the kernel to be copied and then mirrored

Returns:

a reference to this instance

template<class T>

sepKernel<T>& lti::sepKernel< T >::multiply

(

const T &

value

)

multiply each 1D kernel with a constant value

Parameters:

value

the value to be multiplied with

Returns:

a reference to this object

template<class T>

virtual bool lti::sepKernel< T >::read	(	ioHandler &	handler,
		const bool	complete = true
	)			[virtual]

read the object from the given ioHandler

Reimplemented from lti::mathObject.

template<class T>

bool lti::sepKernel< T >::separate	(	const kernel2D< T > &	k,
		const double &	maxDev = 0.01
	)

separate a 2D kernel into 1D kernels

Try to separate the two dimensional kernel k. Stop the separation if the error between original and separated kernel is less than maxDev.

Parameters:

	k	the two dimensional filter to be separated
	maxDev	the maximal deviation per element to be achieved

Returns:

true if the separation succeeded or false otherwise.

template<class T>

void lti::sepKernel< T >::setNorm

(

const T &

newNorm

)

set the norm of each individual 1D kernel to the given value

Parameters:

newNorm

the value to be used as norm

Returns:

a reference to this object

template<class T>

void lti::sepKernel< T >::setNumberOfPairs

(

const int &

numPairs

)

set the number of column/row 1D-filters

template<class T>

T lti::sepKernel< T >::sumOfElements

(

)

const

returns the sum of the elements of the resulting 2D kernel

template<class T>

virtual bool lti::sepKernel< T >::write	(	ioHandler &	handler,
		const bool	complete = true
	)			const [virtual]

write the object in the given ioHandler

Reimplemented from lti::mathObject.

---

Member Data Documentation

template<class T>

std::vector< kernel1D<T> > lti::sepKernel< T >::colKernels [protected]

list of one-dimensional column kernels

Referenced by lti::sepKernel< float >::castFrom(), and lti::sepKernel< float >::getColFilter().

template<class T>

std::vector< kernel1D<T> > lti::sepKernel< T >::rowKernels [protected]

list of one-dimensional row kernels

Referenced by lti::sepKernel< float >::castFrom(), and lti::sepKernel< float >::getRowFilter().

---

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

• ltiLinearKernels.h