FindAverageAndVariance< VALUETYPE > Class Template Reference

Find the average pixel value and its variance in an image or ROI. More...

#include <vigra/inspectimage.hxx>

Public Types
typedef VALUETYPE	argument_type
typedef VALUETYPE	first_argument_type
typedef NumericTraits < VALUETYPE >::RealPromote	result_type
typedef double	second_argument_type
typedef NumericTraits < VALUETYPE >::RealPromote	value_type

Public Member Functions
result_type	average () const
unsigned int	count () const
	FindAverageAndVariance ()
void	operator() (argument_type const &v)
void	operator() (first_argument_type const &v, second_argument_type weight)
void	operator() (FindAverageAndVariance const &v)
result_type	operator() () const
void	reset ()
result_type	variance (bool unbiased=false) const

Detailed Description

template<class VALUETYPE>
class vigra::FindAverageAndVariance< VALUETYPE >

Find the average pixel value and its variance in an image or ROI.

This Functor uses West's algorithm to accumulate highly accurate values for the mean and the sum of squared differences of all values seen so far (the naive incremental algorithm for the computation of the sum of squares produces large round-off errors when the mean is much larger than the standard deviation of the data.) This Functor can also be used in conjunction with ArrayOfRegionStatistics to find the statistics of all regions in a labeled image.

Traits defined:

FunctorTraits::isUnaryAnalyser and FunctorTraits::isInitializer are true (VigraTrueType)

Usage:

#include <vigra/inspectimage.hxx>
Namespace: vigra

vigra::BImage img;

vigra::FindAverageAndVariance<vigra::BImage::PixelType> averageAndVariance;   // init functor

vigra::inspectImage(srcImageRange(img), averageAndVariance);

cout << "Average: " << averageAndVariance.average() << "\n";
cout << "Standard deviation: " << sqrt(averageAndVariance.variance()) << "\n";

Required Interface:

VALUETYPE v1, v2(v1);
double d;

v1 += v2;
v1 + v2;
v1 - v2;
v1 * v2;
v1 / d;
d * v1;

Member Typedef Documentation

typedef VALUETYPE argument_type

the functor's argument type

typedef VALUETYPE first_argument_type

the functor's first argument type (for calls with a weight)

typedef double second_argument_type

the functor's second argument type (for calls with a weight)

typedef NumericTraits<VALUETYPE>::RealPromote result_type

the functor's result type

typedef NumericTraits<VALUETYPE>::RealPromote value_type

Deprecated:

use argument_type and result_type

Constructor & Destructor Documentation

FindAverageAndVariance

(

)

init average

Member Function Documentation

void reset

(

)

(re-)init average and variance

void operator()

(

argument_type const &

v

)

update average and variance

void operator()	(	first_argument_type const &	v,
		second_argument_type	weight
	)

update average and variance, using weighted input. stats(value, 1.0) is equivalent to the unweighted call stats(value), and stats(value, 2.0) is equivalent to two unweighted calls.

void operator()

(

FindAverageAndVariance< VALUETYPE > const &

v

)

merge two statistics

unsigned int count

(

)

const

return number of values (sum of weights) seen so far

result_type average

(

)

const

return current average

result_type variance

(

bool

unbiased = false

)

const

return current variance. If unbiased = true, the sum of squared differences is divided by count()-1 instead of just count().

result_type operator()

(

)

const

return current variance. calls variance().

---

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

• vigra/inspectimage.hxx