Image Residuals

Collaboration diagram for Image Residuals:

Classes
class	ProgImageResiduals

Functions
double	computeCovarianceMatrixDivergence (const Matrix2D< double > &C1, const Matrix2D< double > &C2)
	Compute the divergence between two covariance matrices. More...

Detailed Description

Function Documentation

computeCovarianceMatrixDivergence()

double computeCovarianceMatrixDivergence	(	const Matrix2D< double > &	C1,
		const Matrix2D< double > &	C2
	)

Compute the divergence between two covariance matrices.

Definition at line 99 of file program_image_residuals.cpp.

{
    Matrix1D<double> D;
    Matrix2D<double> P;
    Matrix2D<double> C;
    C=C1+C2;
    C*=0.5;
    firstEigs(C, MAT_XSIZE(C), D, P, false);
    double retval=0;
    for (size_t i=0; i<VEC_XSIZE(D)/2; ++i) // Only half of the eigenvalues are reliable
    {
        double l=fabs(VEC_ELEM(D,i));
        if (l>1e-14)
            retval+=log(l);
    }

    C=C1*C2;
    firstEigs(C, MAT_XSIZE(C), D, P, false);
    for (size_t i=0; i<VEC_XSIZE(D)/2; ++i)
    {
        double l=fabs(VEC_ELEM(D,i));
        if (l>1e-14)
            retval-=0.5*log(l);
    }
    return retval;
}