nmrRegistrationRigid.h File Reference

#include <cisstCommon/cmnPortability.h>
#include <cisstVector/vctTypes.h>
#include <cisstNumerical/nmrExport.h>

Go to the source code of this file.

Functions
template<class _vectorOwnerType>
bool CISST_EXPORT	nmrRegistrationRigid (vctDynamicConstVectorBase< _vectorOwnerType, vct3 > &dataSet1, vctDynamicConstVectorBase< _vectorOwnerType, vct3 > &dataSet2, vctFrm3 &transform, double *fre=0)

Function Documentation

nmrRegistrationRigid()

template<class _vectorOwnerType>

bool CISST_EXPORT nmrRegistrationRigid	(	vctDynamicConstVectorBase< _vectorOwnerType, vct3 > &	dataSet1,
		vctDynamicConstVectorBase< _vectorOwnerType, vct3 > &	dataSet2,
		vctFrm3 &	transform,
		double *	fre = 0 )

Perform a paired-point rigid registration using the SVD method proposed by Arun(1987) and modified by Umeyama(1991).

Parameters

dataSet1	The first set of fiducial points
dataSet2	The second set of fiducial points
transform	The computed transformation from dataSet1 to dataSet2
fre	Pointer to location to store Fiducial Registration Error (FRE), if not 0.

Returns

true if registration successful

Note

This function is templated so that it will work when the data sets are of type vctDynamicConstVectorRef or vctDynamicConstVector. The explicit template instantiations are in the cpp file.

This function is wrapped for Python, as follows:

(transform, fre) = nmrRigidRegistration(dataSet1, dataSet2)

where dataSet1 and dataSet2 are 2D numpy arrays, such as ((1.0, 2.0, 3.0), (4.0, 5.0, 6.0), ...). The Python function throws an exception if registration fails for any reason, including problems with the input data (e.g., data sets not the same number of points).