Cvb::Foundation::Analysis Namespace Reference

Namespace for collection of miscellaneous image analysis functions. More...

Functions
void	LocalEigenValuesAndVectors (const ImagePlane &plane, Filter::FixedFilterSize derivatorSize, Filter::FixedFilterSize blurSize, std::unique_ptr< Image > &lambda1, std::unique_ptr< Image > &lambda2, std::unique_ptr< Image > &eigenValue1x, std::unique_ptr< Image > &eigenValue1y, std::unique_ptr< Image > &eigenValue2x, std::unique_ptr< Image > &eigenValue2y)
	Calculate the local eigenvalues and eigenvectors of the Hesse matrix applied to the pixels of the input image.
std::unique_ptr< Image >	LocalMinEigenValues (const ImagePlane &plane, Filter::FixedFilterSize derivatorSize, Filter::FixedFilterSize blurSize)
	Calculate the minimum local eigenvalues of the Hesse matrix applied to the pixels of the input image.
double	WangQuality (const ImagePlane &plane1, const ImagePlane &plane2)
	Calculate the Wang quality for two images.

Detailed Description

Namespace for collection of miscellaneous image analysis functions.

Remarks

CMake users: Link to imported target CVB::CvbFoundationAnalysis

Function Documentation

LocalEigenValuesAndVectors()

void LocalEigenValuesAndVectors ( const ImagePlane & plane, Filter::FixedFilterSize derivatorSize, Filter::FixedFilterSize blurSize, std::unique_ptr< Image > & lambda1, std::unique_ptr< Image > & lambda2, std::unique_ptr< Image > & eigenValue1x, std::unique_ptr< Image > & eigenValue1y, std::unique_ptr< Image > & eigenValue2x, std::unique_ptr< Image > & eigenValue2y )

inline

Calculate the local eigenvalues and eigenvectors of the Hesse matrix applied to the pixels of the input image.

Parameters

[in]	plane	Image plane to work on.
[in]	derivatorSize	Size of the derivation operator (Sobel, valid inputs are Filter::FixedFilterSize::Kernel3x3 and Filter::FixedFilterSize::Kernel5x5).
[in]	blurSize	Size of the blurring operator (valid inputs are Filter::FixedFilterSize::Kernel3x3 and Filter::FixedFilterSize::Kernel5x5).
[out]	lambda1	First eigenvalue for each pixel (output).
[out]	lambda2	Second eigenvalue for each pixel (output).
[out]	eigenValue1x	X-component of the first eigenvector (output).
[out]	eigenValue1y	Y-component of the first eigenvector (output).
[out]	eigenValue2x	X-component of the second eigenvector (output).
[out]	eigenValue2y	Y-component of the second eigenvector (output).

Exceptions

Any	exception derived from std::exception including CvbException.

LocalMinEigenValues()

std::unique_ptr< Image > LocalMinEigenValues ( const ImagePlane & plane, Filter::FixedFilterSize derivatorSize, Filter::FixedFilterSize blurSize )

inline

Calculate the minimum local eigenvalues of the Hesse matrix applied to the pixels of the input image.

Parameters

[in]	plane	Image plane to work on.
[in]	derivatorSize	Size of the derivation operator (Sobel, valid inputs are Filter::FixedFilterSize::Kernel3x3 and Filter::FixedFilterSize::Kernel5x5).
[in]	blurSize	Size of the blurring operator (valid inputs are Filter::FixedFilterSize::Kernel3x3 and Filter::FixedFilterSize::Kernel5x5).

Returns

Image with the minimal local eigenvaluesThe filtered image

Exceptions

Any	exception derived from std::exception including CvbException.

This approach helps identify locations inside the image that contain curved edges.

WangQuality()

double WangQuality ( const ImagePlane & plane1, const ImagePlane & plane2 )

inline

Calculate the Wang quality for two images.

Parameters

[in]	plane1	Image plane 1 on which to calculate.
[in]	plane2	Image plane 2 on which to calculate.

Returns

The quality index (ranging from -1 to 1)

Exceptions

Any	exception derived from std::exception including CvbException.

The Wang Quality is a measure of similarity between two images as proposed in a publication by Z. Wang and A. C. Bovik (IEEE Signal Processing Letters, vol. 9, no. 3, pp. 81-84, March 2002).