This class is a specialized WeightedLinearFilter used to retouch missing data values. More...

#include <filtering.hpp>

Inheritance diagram for NumeRe::RetouchRegion:

Collaboration diagram for NumeRe::RetouchRegion:

Public Member Functions
	RetouchRegion (size_t _row, size_t _col, const mu::value_type &_dMedian)
	Filter constructor. Will automatically create the filter kernel. More...

virtual	~RetouchRegion () override
	Filter destructor. Will clear the previously calculated filter kernel. More...

virtual double	operator() (size_t i, size_t j) const override
	Override for the operator(). Returns the filter kernel at the desired position. More...

virtual mu::value_type	apply (size_t i, size_t j, const mu::value_type &val) const override
	Override for the abstract apply method of the base class. Applies the filter to the value at the selected position and returns the new value. More...

void	setBoundaries (const std::vector< mu::value_type > &left, const std::vector< mu::value_type > &right, const std::vector< mu::value_type > &top=std::vector< mu::value_type >(), const std::vector< mu::value_type > &bottom=std::vector< mu::value_type >())
	This method is used to update the internal filter boundaries. More...

mu::value_type	retouch (size_t i, size_t j, const mu::value_type &val, const mu::value_type &med)
	This method is a wrapper to retouch only invalid values. The default value of invalid values is the median value declared at construction time. More...

Public Member Functions inherited from NumeRe::Filter
	Filter (size_t row, size_t col)
	Filter base constructor. Will set the used window sizes. More...

virtual	~Filter ()
	Empty virtual abstract destructor. More...

virtual double	operator() (size_t i, size_t j) const =0
	Virtual operator() override. Has to be implemented in the child classes and shall return the kernel value at position (i,j). More...

virtual mu::value_type	apply (size_t i, size_t j, const mu::value_type &val) const =0
	Virtual method for applying the filter to a distinct value. Has to be implemented in all child classes. More...

bool	isConvolution () const
	This method returns, whether the current filter is a convolution, ie. whether the returned value may be used directly or if all values of a window have to be accumulated first. More...

FilterSettings::FilterType	getType () const
	This method returns the type of the current filter as a value of the FilterType enumeration. More...

std::pair< size_t, size_t >	getWindowSize () const
	This method returns the window size of the current filter as a std::pair in the order (row,col). More...

FilterBuffer &	getBuffer ()
	This method returns the internal filtering buffer queue to store already smoothed points avoiding leakage effects. More...

FilterBuffer2D &	get2DBuffer ()
	This method returns the internal filtering buffer queue for 2D data to store already smoothed points avoiding leakage effects. More...

Private Member Functions
void	createKernel ()
	This method will create the filter's kernel for the selected window size. More...

mu::value_type	left (size_t i, size_t j) const
	This method will return the correct value for the left interval boundary. More...

mu::value_type	right (size_t i, size_t j) const
	This method will return the correct value for the right interval boundary. More...

mu::value_type	top (size_t i, size_t j) const
	This method will return the correct value for the top interval boundary. More...

mu::value_type	bottom (size_t i, size_t j) const
	This method will return the correct value for the bottom interval boundary. More...

mu::value_type	topleft (size_t i, size_t j) const
	This method will return the correct value for the topleft diagonal interval boundary. More...

mu::value_type	topright (size_t i, size_t j) const
	This method will return the correct value for the topright diagonal interval boundary. More...

mu::value_type	bottomleft (size_t i, size_t j) const
	This method will return the correct value for the bottomleft diagonal interval boundary. More...

mu::value_type	bottomright (size_t i, size_t j) const
	This method will return the correct value for the bottomright diagonal interval boundary. More...

mu::value_type	validize (mu::value_type val) const
	This method checks, whether the passed value is a valid value and returns it. If it is not, it will be replaced by the fallback value. More...

Private Attributes
std::vector< mu::value_type >	m_left

std::vector< mu::value_type >	m_right

std::vector< mu::value_type >	m_top

std::vector< mu::value_type >	m_bottom

bool	is2D

std::vector< std::vector< double > >	m_filterKernel

mu::value_type	m_fallback

bool	m_invertedKernel

Additional Inherited Members
Protected Attributes inherited from NumeRe::Filter
FilterSettings::FilterType	m_type

std::pair< size_t, size_t >	m_windowSize

bool	m_isConvolution

FilterBuffer	m_buffer

FilterBuffer2D	m_buffer2D

Detailed Description

This class is a specialized WeightedLinearFilter used to retouch missing data values.

Definition at line 837 of file filtering.hpp.

Constructor & Destructor Documentation

◆ RetouchRegion()

NumeRe::RetouchRegion::RetouchRegion	(	size_t	_row,
		size_t	_col,
		const mu::value_type &	_dMedian
	)

inline

Filter constructor. Will automatically create the filter kernel.

Parameters

_row	size_t
_col	size_t
_dMedian	const mu::value_type&

Definition at line 1047 of file filtering.hpp.

References createKernel(), NumeRe::FilterSettings::FILTER_WEIGHTED_LINEAR, is2D, m_fallback, m_invertedKernel, NumeRe::Filter::m_isConvolution, and NumeRe::Filter::m_type.

Here is the call graph for this function:

◆ ~RetouchRegion()

virtual NumeRe::RetouchRegion::~RetouchRegion ( )

inlineoverridevirtual

Filter destructor. Will clear the previously calculated filter kernel.

Definition at line 1062 of file filtering.hpp.

References m_filterKernel.

Member Function Documentation

◆ apply()

virtual mu::value_type NumeRe::RetouchRegion::apply	(	size_t	i,
		size_t	j,
		const mu::value_type &	val
	)		const

inlineoverridevirtual

Override for the abstract apply method of the base class. Applies the filter to the value at the selected position and returns the new value.

Parameters

i	size_t
j	size_t
val	const mu::value_type&

Returns: virtual mu::value_type

Implements NumeRe::Filter.

Definition at line 1096 of file filtering.hpp.

References bottom(), bottomleft(), bottomright(), is2D, left(), m_filterKernel, m_invertedKernel, NumeRe::Filter::m_windowSize, right(), top(), topleft(), and topright().

Referenced by retouch().

Here is the call graph for this function:

◆ bottom()

mu::value_type NumeRe::RetouchRegion::bottom	(	size_t	i,
		size_t	j
	)		const

inlineprivate

This method will return the correct value for the bottom interval boundary.

Parameters

i	size_t
j	size_t

Returns: mu::value_type

Definition at line 939 of file filtering.hpp.

References is2D, m_bottom, and validize().

Referenced by apply(), and setBoundaries().

Here is the call graph for this function:

◆ bottomleft()

mu::value_type NumeRe::RetouchRegion::bottomleft	(	size_t	i,
		size_t	j
	)		const

inlineprivate

This method will return the correct value for the bottomleft diagonal interval boundary.

Parameters

i	size_t
j	size_t

Returns: mu::value_type

Definition at line 993 of file filtering.hpp.

References m_bottom, m_left, NumeRe::Filter::m_windowSize, and validize().

Referenced by apply().

Here is the call graph for this function:

◆ bottomright()

mu::value_type NumeRe::RetouchRegion::bottomright	(	size_t	i,
		size_t	j
	)		const

inlineprivate

This method will return the correct value for the bottomright diagonal interval boundary.

Parameters

i	size_t
j	size_t

Returns: mu::value_type

Definition at line 1011 of file filtering.hpp.

References m_bottom, m_right, NumeRe::Filter::m_windowSize, and validize().

Referenced by apply().

Here is the call graph for this function:

◆ createKernel()

void NumeRe::RetouchRegion::createKernel ( )

inlineprivate

This method will create the filter's kernel for the selected window size.

Returns: void

Definition at line 856 of file filtering.hpp.

References is2D, m_filterKernel, NumeRe::Filter::m_windowSize, max, and NumeRe::pow2().

Referenced by RetouchRegion().

Here is the call graph for this function:

◆ left()

mu::value_type NumeRe::RetouchRegion::left	(	size_t	i,
		size_t	j
	)		const

inlineprivate

This method will return the correct value for the left interval boundary.

Parameters

i	size_t
j	size_t

Returns: mu::value_type

Definition at line 888 of file filtering.hpp.

References is2D, m_left, and validize().

Referenced by apply(), and setBoundaries().

Here is the call graph for this function:

◆ operator()()

virtual double NumeRe::RetouchRegion::operator()	(	size_t	i,
		size_t	j
	)		const

inlineoverridevirtual

Override for the operator(). Returns the filter kernel at the desired position.

Parameters

i	size_t
j	size_t

Returns: double

Implements NumeRe::Filter.

Definition at line 1076 of file filtering.hpp.

References m_filterKernel, and NumeRe::Filter::m_windowSize.

◆ retouch()

mu::value_type NumeRe::RetouchRegion::retouch	(	size_t	i,
		size_t	j,
		const mu::value_type &	val,
		const mu::value_type &	med
	)

inline

This method is a wrapper to retouch only invalid values. The default value of invalid values is the median value declared at construction time.

Parameters

i	size_t
j	size_t
val	const mu::value_type&
med	const mu::value_type&

Returns: mu::value_type

Definition at line 1172 of file filtering.hpp.

References apply(), mu::isnan(), and m_fallback.

Referenced by Memory::retouch2D().

Here is the call graph for this function:

◆ right()

mu::value_type NumeRe::RetouchRegion::right	(	size_t	i,
		size_t	j
	)		const

inlineprivate

This method will return the correct value for the right interval boundary.

Parameters

i	size_t
j	size_t

Returns: mu::value_type

Definition at line 905 of file filtering.hpp.

References is2D, m_right, and validize().

Referenced by apply(), and setBoundaries().

Here is the call graph for this function:

◆ setBoundaries()

void NumeRe::RetouchRegion::setBoundaries	(	const std::vector< mu::value_type > &	left,
		const std::vector< mu::value_type > &	right,
		const std::vector< mu::value_type > &	top = `std::vector<mu::value_type>()`,
		const std::vector< mu::value_type > &	bottom = `std::vector<mu::value_type>()`
	)

inline

This method is used to update the internal filter boundaries.

Parameters

left	const std::vector<mu::value_type>&
right	const std::vector<mu::value_type>&
top	const std::vector<mu::value_type>&
bottom	const std::vector<mu::value_type>&