Common useful measures on images, not based on Morphological operations. More...

Detailed Description

Common useful measures on images, not based on Morphological operations.

Collaboration diagram for Base measures:

Functions
template<class T >
size_t	area (const Image< T > &imIn)
	area() - Area of an image

template<class T >
double	vol (const Image< T > &imIn)
	vol() - Volume of an image

template<class T >
double	volume (const Image< T > &imIn)
	colume() - Volume of an image

template<class T >
Vector_double	meanVal (const Image< T > &imIn, bool onlyNonZero=false)
	meanVal() - Mean value and standard deviation

template<class T >
T	minVal (const Image< T > &imIn, bool onlyNonZero=false)
	minVal() - Min value of an image

template<class T >
T	minVal (const Image< T > &imIn, Point< UINT > &pt, bool onlyNonZero=false)
	minVal() - Min value of an image

template<class T >
T	maxVal (const Image< T > &imIn, bool onlyNonZero=false)
	maxVal() - Max value of an image

template<class T >
T	maxVal (const Image< T > &imIn, Point< UINT > &pt, bool onlyNonZero=false)
	maxVal() - Max value of an image

template<class T >
vector< T >	rangeVal (const Image< T > &imIn, bool onlyNonZero=false)
	rangeVal() - Min and Max values of an image

template<class T >
vector< T >	valueList (const Image< T > &imIn, bool onlyNonZero=true)
	valueList() - Get the list of the pixel values present in the image

template<class T >
T	modeVal (const Image< T > &imIn, bool onlyNonZero=true)
	modeVal() - Get the mode of the histogram present in the image, i.e.

template<class T >
T	medianVal (const Image< T > &imIn, bool onlyNonZero=true)
	medianVal() - Get the median of the image histogram.

template<class T >
vector< T >	profile (const Image< T > &im, size_t x0, size_t y0, size_t x1, size_t y1, size_t z=0)
	profile() - Get image values along a line defined by the points and in the slice .

template<class T >
Vector_double	measBarycenter (Image< T > &im)
	measBarycenter() - Gets the barycenter coordinates of an image

template<class T >
vector< size_t >	measBoundBox (Image< T > &im)
	measBoundBox() - Bounding Box measure - gets the coordinates of the bounding box

template<class T >
Vector_double	measMoments (Image< T > &im, const bool onlyNonZero=true, const bool centered=false)
	measMoments() - Measure image moments

template<class T >
vector< double >	measCovariance (const Image< T > &imIn1, const Image< T > &imIn2, size_t dx, size_t dy, size_t dz, size_t maxSteps=0, bool centered=false, bool normalize=false)
	measCovariance() - Centered covariance of two images in the direction defined by dx, dy and dz.

template<class T >
vector< double >	measAutoCovariance (const Image< T > &imIn, size_t dx, size_t dy, size_t dz, size_t maxSteps=0, bool centered=false, bool normalize=false)
	measAutoCovariance() - Auto-covariance

template<class T >
double	measEntropy (const Image< T > &imIn)
	measEntropy() - Image entropy

template<class T >
double	measEntropy (const Image< T > &imIn, const Image< T > &imMask)
	measEntropy() - Image entropy

template<class T >
Vector_size_t	nonZeroOffsets (Image< T > &imIn)
	nonZeroOffsets() - Returns the offsets of pixels having non nul values.

template<class T >
bool	isBinary (const Image< T > &imIn)
	isBinary() - Test if an image is binary.

Function Documentation

◆ area()

template<class T >

size_t area ( const Image< T > & imIn )

area() - Area of an image

The area of an image is defined as the number of non-zero pixels

Parameters

[in] imIn : Input image.

Returns: the number of non-zero pixels

Note: The name of this function comes from times where images were mostly 2D only. In those days, the area of an image was said to be the area of xy plane where pixel values are non-zero, i.e., the number of non-zero pixels. This may be confusing for 3D images, but the idea remains the same.

◆ vol()

template<class T >

double vol ( const Image< T > & imIn )

vol() - Volume of an image

Parameters

[in] imIn : Input image.

Returns: the sum of pixel values

This is the same than the volume function call. Better use the unabridged name. The abridged name remains for back compatibility.

◆ volume()

template<class T >

double volume ( const Image< T > & imIn )

colume() - Volume of an image

The volume of an image is defined as the sum of the pixel values.

Parameters

[in] imIn : Input image.

Returns: the sum of pixel values (as a double)

Note: The name of this function comes from times where images were mostly 2D only. In those days, the volume of an image was said to be the volume defined by the xy plane with the third dimension being the intensity (pixel values). This may be confusing for 3D images, but the idea remains the same.

◆ meanVal()

template<class T >

Vector_double meanVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `false`
	)

meanVal() - Mean value and standard deviation

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: a vector with the mean and standard deviation of pixel values

◆ minVal() [1/2]

template<class T >

T minVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `false`
	)

minVal() - Min value of an image

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: the min of the pixel values.

◆ minVal() [2/2]

template<class T >

T minVal	(	const Image< T > &	imIn,
		Point< UINT > &	pt,
		bool	onlyNonZero = `false`
	)

minVal() - Min value of an image

Parameters

[in]	imIn	: Input image.
[out]	pt	: point coordinates of the minimum value in the image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: the min of the pixel values.

◆ maxVal() [1/2]

template<class T >

T maxVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `false`
	)

maxVal() - Max value of an image

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: the max of the pixel values.

◆ maxVal() [2/2]

template<class T >

T maxVal	(	const Image< T > &	imIn,
		Point< UINT > &	pt,
		bool	onlyNonZero = `false`
	)

maxVal() - Max value of an image

Parameters

[in]	imIn	: Input image.
[out]	pt	: point coordinates of the maximum value in the image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: the max of the pixel values.

◆ rangeVal()

template<class T >

vector< T > rangeVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `false`
	)

rangeVal() - Min and Max values of an image

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: a vector with the min and the max of the pixel values.

◆ valueList()

template<class T >

vector< T > valueList	(	const Image< T > &	imIn,
		bool	onlyNonZero = `true`
	)

valueList() - Get the list of the pixel values present in the image

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: a vector with the values found in the image.

See also: histogram

Warning: In huge images of type UINT32, this function may return a huge vector.

◆ modeVal()

template<class T >

T modeVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `true`
	)

modeVal() - Get the mode of the histogram present in the image, i.e.

the value that appears most often.

Parameters

[in]	imIn	: input image
[in]	onlyNonZero	: consider only non zero values

Returns: the value that appears more often

Note

As this function returns only one value :

in a distribution with the same maximum for many values, it returns the first one;
in a multimodal distribution, it returns the first biggest one;

◆ medianVal()

template<class T >

T medianVal	(	const Image< T > &	imIn,
		bool	onlyNonZero = `true`
	)

medianVal() - Get the median of the image histogram.

Parameters

[in]	imIn	: Input image.
[in]	onlyNonZero	: If true, only non-zero pixels are considered.

Returns: the median of the image histogram

◆ profile()

template<class T >

vector< T > profile	(	const Image< T > &	im,
		size_t	x0,
		size_t	y0,
		size_t	x1,
		size_t	y1,
		size_t	z = `0`
	)

profile() - Get image values along a line defined by the points and in the slice .

Parameters

[in]	im	: input image
[in]	x0,y0	: start point
[in]	x1,y1	: end point
[in]	z	: slice

Returns: vector with pixel values

◆ measBarycenter()

template<class T >

Vector_double measBarycenter ( Image< T > & im )

measBarycenter() - Gets the barycenter coordinates of an image

Parameters

[in] im : input image

Returns: vector with the coordinates of barycenter

◆ measBoundBox()

template<class T >

vector< size_t > measBoundBox ( Image< T > & im )

measBoundBox() - Bounding Box measure - gets the coordinates of the bounding box

Parameters

[in] im : input image

Returns: a vector with xMin, yMin (,zMin), xMax, yMax (,zMax)

◆ measMoments()

template<class T >

Vector_double measMoments	(	Image< T > &	im,
		const bool	onlyNonZero = `true`,
		const bool	centered = `false`
	)

measMoments() - Measure image moments

Parameters

[in]	im	: input image
[in]	onlyNonZero	: use only non zero values
[in]	centered	: returns centered moments

Returns: For 2D images: vector(m00, m10, m01, m11, m20, m02); For 3D images: vector(m000, m100, m010, m001, m110, m101, m011, m200, m020, m002)

See also: blobsMoments() call if you want to evaluate moments for each blob.; Image moment on Wikipedia

Inertia matrix can be evaluated :

For 3D images :
$M = \begin{bmatrix} m020 + m002 & -m110 & -m101 \\ -m110 & m200 + m002 & -m011 \\ -m101 & -m011 & m200 + m020 \end{bmatrix}$
For 2D images :
$M = \begin{bmatrix} m20 & -m11 \\ -m11 & m02 \end{bmatrix}$

◆ measCovariance()

template<class T >

vector< double > measCovariance	(	const Image< T > &	imIn1,
		const Image< T > &	imIn2,
		size_t	dx,
		size_t	dy,
		size_t	dz,
		size_t	maxSteps = `0`,
		bool	centered = `false`,
		bool	normalize = `false`
	)

measCovariance() - Centered covariance of two images in the direction defined by dx, dy and dz.

The direction is given by dx, dy and dz.

The lenght corresponds to the max number of steps maxSteps. When 0, the length is limited by the dimensions of the image.

$vec[h] = \sum_{p \:\in\: imIn1} \frac{imIn1(p) \;.\; imIn2(p + h)}{N_p}$

where h are displacements in the direction defined by dx, dy and dz.

is the number of pixels used in each term of the sum, which may different for each term in the sum.

Parameters

[in]	imIn1,imIn2	: Input Images
[in]	dx,dy,dz	: direction
[in]	maxSteps	: number maximum of displacements to evaluate
[in]	centered	: if this parameter is set to true, the mean value (meanVal()) will be subtracted from each input image
[in]	normalize	: normalize result with respect to vec[0]

Returns: vec[h]

◆ measAutoCovariance()

template<class T >

vector< double > measAutoCovariance	(	const Image< T > &	imIn,
		size_t	dx,
		size_t	dy,
		size_t	dz,
		size_t	maxSteps = `0`,
		bool	centered = `false`,
		bool	normalize = `false`
	)

measAutoCovariance() - Auto-covariance

The direction is given by dx, dy and dz. The lenght corresponds to the max number of steps maxSteps

Parameters

[in]	imIn	: Input Image
[in]	dx,dy,dz	: direction
[in]	maxSteps	: number maximum of displacements to evaluate
[in]	centered	: if this parameter is set to true, the mean value (meanVal()) will be subtracted from the input image
[in]	normalize	: normalize result with respect to vec[0]

Returns: vec[h]

◆ measEntropy() [1/2]

template<class T >

double measEntropy ( const Image< T > & imIn )

measEntropy() - Image entropy

Evaluate Shannon entropy of the image (in bits)

Parameters

[in] imIn : input image

Returns: image entropy

See also: blobsEntropy() to compute entropy inside each label.

◆ measEntropy() [2/2]

template<class T >

double measEntropy	(	const Image< T > &	imIn,
		const Image< T > &	imMask
	)

measEntropy() - Image entropy

Evaluate Shannon entropy of the image in a region defined by a mask.

Parameters

[in]	imIn	: input image
[in]	imMask	: mask defining where the entropy shall be evaluated

Returns: image entropy

See also: blobsEntropy() to compute entropy inside each label.

◆ nonZeroOffsets()

template<class T >

Vector_size_t nonZeroOffsets ( Image< T > & imIn )

nonZeroOffsets() - Returns the offsets of pixels having non nul values.

Parameters

[in] imIn : input image

Returns: a vector containing the offset of all non-zero points in image.

Warning: In huge images this can return a very big vector, in the same order than the image size.

◆ isBinary()

template<class T >

bool isBinary ( const Image< T > & imIn )

isBinary() - Test if an image is binary.

Parameters

[in] imIn : image

Returns: true if the only pixel values are 0 and any other positive value.

Detailed Description

Functions

Function Documentation

◆ area()

◆ vol()

◆ volume()

◆ meanVal()

◆ minVal() [1/2]

◆ minVal() [2/2]

◆ maxVal() [1/2]

◆ maxVal() [2/2]

◆ rangeVal()

◆ valueList()

◆ modeVal()

◆ medianVal()

◆ profile()

◆ measBarycenter()

◆ measBoundBox()

◆ measMoments()

◆ measCovariance()

◆ measAutoCovariance()

◆ measEntropy() [1/2]

◆ measEntropy() [2/2]

◆ nonZeroOffsets()

◆ isBinary()