#include <MedianFilter.h>

Public Member Functions
	MedianFilter (int box_width, int box_height)

	MedianFilter (const std::array< int, 2 > &box)

auto	operator() (const VectorImage< T > &image, const VectorImage< T > &variance, T threshold=0) const -> std::pair< std::shared_ptr< VectorImage< T > >, std::shared_ptr< VectorImage< T > > >

Private Member Functions
std::vector< T >	getBox (const VectorImage< T > &img, int x, int y) const

Static Private Member Functions
static T	getMedian (std::vector< T > &data)

static int	clip (int position, int box_size, int image_size)

Private Attributes
int	m_box_width

int	m_box_height

Detailed Description

template<typename T>
class SourceXtractor::MedianFilter< T >

Functor that applies a median filter over a pair of image/variance image

Template Parameters

T	Pixel type of the filtered image

A median filter takes as parameters the width and height of the smoothing box. Each pixel is computed as the median of the values of the pixels contained within a box centered at its coordinates iff the median value differs from the actual value by more than a given threshold.

If the value of the pixel is replaced, the value of the variance at the same coordinates is replaced also by the median of the variances contained within the box.

There is no padding. At the edges whe box size is symmetrically clipped to the number of available pixels.

Definition at line 41 of file MedianFilter.h.

Constructor & Destructor Documentation

◆ MedianFilter() [1/2]

template<typename T >

SourceXtractor::MedianFilter< T >::MedianFilter	(	int	box_width,
		int	box_height )

inline

Constructor

Parameters

box_width
box_height

Definition at line 48 of file MedianFilter.h.

◆ MedianFilter() [2/2]

template<typename T >

SourceXtractor::MedianFilter< T >::MedianFilter ( const std::array< int, 2 > & box )

inlineexplicit

Constructor

Parameters

box	A two dimensional array where the first value corresponds to the width, and the second to the height.

Definition at line 56 of file MedianFilter.h.

Member Function Documentation

◆ clip()

template<typename T >

static int SourceXtractor::MedianFilter< T >::clip	(	int	position,
		int	box_size,
		int	image_size )

inlinestaticprivate

Symmetrically clip the size of the smoothing box depending on the number of available pixels. This is a 1-D version that is applied to X and Y separately.

Parameters

position	Coordinate where the box will be centered
box_size	Size of the smoothing box
image_size	Image size

Returns: As many pixels as could be read safely from the image, up to box_size

Definition at line 127 of file MedianFilter.h.

Referenced by SourceXtractor::MedianFilter< T >::getBox().

◆ getBox()

template<typename T >

std::vector< T > SourceXtractor::MedianFilter< T >::getBox	(	const VectorImage< T > &	img,
		int	x,
		int	y ) const

inlineprivate

Get the pixel values contained within a box centered at x,y

Definition at line 139 of file MedianFilter.h.

References std::back_inserter(), SourceXtractor::MedianFilter< T >::clip(), std::inserter(), SourceXtractor::MedianFilter< T >::m_box_height, SourceXtractor::MedianFilter< T >::m_box_width, x, and y.

Referenced by SourceXtractor::MedianFilter< T >::operator()().

Here is the call graph for this function:

◆ getMedian()

template<typename T >

static T SourceXtractor::MedianFilter< T >::getMedian ( std::vector< T > & data )

inlinestaticprivate

Convenience method to compute the median of a vector

Parameters

data	The vector on which to compute the median. It will be modified.

Definition at line 107 of file MedianFilter.h.

References std::sort().

Referenced by SourceXtractor::MedianFilter< T >::operator()().

Here is the call graph for this function:

◆ operator()()

template<typename T >

auto SourceXtractor::MedianFilter< T >::operator()	(	const VectorImage< T > &	image,
		const VectorImage< T > &	variance,
		T	threshold = 0 ) const -> std::pair<std::shared_ptr<VectorImage<T>>, std::shared_ptr<VectorImage<T>>>

inline

Call operator

Parameters

image	Image to filter
variance	Associated variance map
threshold	The pixel is replaced only if the median and the actual value differ by at least this value. Defaults to 0 (always replaced).