IMP::em::CoarseCC Class Reference

Detailed Description

Responsible for performing coarse fitting between two density objects.

The pixels involved are derived from the positions of N particles.

Inheritance diagram for IMP::em::CoarseCC:


Static Public Member Functions
static void	calc_derivatives (const DensityMap &em_map, SampledDensityMap &model_map, const float &scalefac, std::vector< float > &dvx, std::vector< float > &dvy, std::vector< float > &dvz)
static float	cross_correlation_coefficient (const DensityMap &em_map, DensityMap &model_map, float voxel_data_threshold, bool recalc_ccnormfac=true, bool divide_by_rms=true)
	Calculates the cross correlation coefficient between two maps.
Related Functions
(Note that these are not member functions.)
static float	evaluate (DensityMap &data, SampledDensityMap &model_map, std::vector< float > &dvx, std::vector< float > &dvy, std::vector< float > &dvz, float scalefactor, bool lderiv, bool divide_by_rms=true, bool resample=true)
	Evaluates the value of the EM fitting term.

Member Function Documentation

static void IMP::em::CoarseCC::calc_derivatives	(	const DensityMap &	em_map,
		SampledDensityMap &	model_map,
		const float &	scalefac,
		std::vector< float > &	dvx,
		std::vector< float > &	dvy,
		std::vector< float > &	dvz
	)			`[static]`

Computes the derivatives of the cross correlation term scalefac*(1-ccc) at each voxel of the map.

Parameters:

`[in]`	em_map	the target density map.
`[in]`	model_map	the sampled density map of the model
`[in]`	scalefac	scale factor to apply to the value of the cross correlation term
`[out]`	dvx	vector to contain the x partial derivatives
`[out]`	dvy	vector to contain the y partial derivatives
`[out]`	dvz	vector to contain the z partial derivatives

Note:: : The function assumes that correct RMS are calculated for the densities

static float IMP::em::CoarseCC::cross_correlation_coefficient	(	const DensityMap &	em_map,
		DensityMap &	model_map,
		float	voxel_data_threshold,
		bool	recalc_ccnormfac = `true`,
		bool	divide_by_rms = `true`
	)			`[static]`

Calculates the cross correlation coefficient between two maps.

Cross correlation coefficient between the em density and the density of a model. The function applied is: $\frac{\sum_{i=1}^{N}{{td}_i}{{md}_i}-{N} {{mean}_{td}} {{mean}_{md}}} {N\sigma_{{td}}\sigma_{{md}}}$ , such that $N$ is the number of voxels, ${td}$ is the target density, ${tm}$ is the model density,

Parameters:

`[in]`	em_map	the target map (experimentally determined)
`[in]`	model_map	the sampled density map of the model
`[in]`	voxel_data_threshold	voxels with value lower than threshold in model_map are not summed (avoid calculating correlation on voxels below the threshold)
`[in]`	recalc_ccnormfac	determines wheather the model_map should be normalized prior to the correlation calculation. false is faster, but potentially innacurate
`[in]`	divide_by_rms	determines wheather the model_map should be normalized prior to the correlation calculation. false is faster, but potentially innacurate

Returns:: the cross correlation coefficient value between two density maps

Note:: This is not the local CC function

Friends And Related Function Documentation

static float evaluate	(	DensityMap &	data,
		SampledDensityMap &	model_map,
		std::vector< float > &	dvx,
		std::vector< float > &	dvy,
		std::vector< float > &	dvz,
		float	scalefactor,
		bool	lderiv,
		bool	divide_by_rms = `true`,
		bool	resample = `true`
	)			`[related]`

Evaluates the value of the EM fitting term.

Note:: The function returns scalefac*(1-ccc) to support minimization optimization. The ccc value (cross correlation coefficient) is calculate by the cross_correlation_coefficient function.

Parameters:

`[in]`	data	DensityMap class containing the EM map. note: correct RMSD and mean MUST be in the header!
`[in]`	model_map	SampledDensityMap class prepared to contain the simulated EM map for the model.
`[in]`	dvx	vector to contain the xpartial derivatives
`[in]`	dvy	vector to contain the y partial derivatives
`[in]`	dvz	vector to contain the z partial derivatives
`[in]`	scalefactor	scale factor to apply to the value of the cross correlation term
`[in]`	lderiv	if true, the derivatives of the term are computed
`[in]`	divide_by_rms	determines wheather the model_map should be normalized prior to the correlation calculation. false is faster, but potentially innacurate
`[in]`	resample	if true, the model density map is resampled

Returns:: the value of the cross correlation term: scalefac*(1-ccc)

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

CoarseCC.h
em/__init__.py

IMP::em::CoarseCC Class Reference

Detailed Description

Static Public Member Functions

Related Functions

Member Function Documentation

Friends And Related Function Documentation