#include <irtkImageEigenFreeFormRegistration.h>
This class implements a registration filter for the adaptive hierarchical non-rigid registration of two images by manipulating the shape parameter vector of the eigen system rather than the control points. This algorithm is yet to be published.
Definition at line 30 of file irtkImageEigenFreeFormRegistration.h.
Public Member Functions | |
irtkImageEigenFreeFormRegistration () | |
Constructor. | |
virtual | ~irtkImageEigenFreeFormRegistration () |
Destructor. | |
virtual | SetMacro (NumberOfModes, int) |
virtual | GetMacro (NumberOfModes, int) |
virtual const char * | NameOfClass () |
Returns the name of the class. | |
virtual double | EvaluateGradient (float, float *) |
Evaluates the gradient of the similarity metric. | |
Protected Member Functions | |
virtual void | Initialize (int) |
Initial set up for the registration at a multiresolution level. | |
virtual void | Finalize (int) |
Final set up for the registration at a multiresolution level. | |
Protected Attributes | |
int | _NumberOfModes |
Number of modes. | |
irtkEigenFreeFormTransformation * | _effd |
Short cut to local transformation (includes eigen vectors and values). |
virtual double irtkImageEigenFreeFormRegistration::EvaluateGradient | ( | float | , | |
float * | ||||
) | [virtual] |
Evaluates the gradient of the similarity metric.
This function evaluates the gradient of the similarity metric of the registration by looping over the target image and interpolating the transformed source image while filling the joint histogram. The partial derivatives are approximated using a finite difference scheme. The step size for the finite difference scheme is passed as a parameter to the function. The function returns the norm of the gradient vector as well as the gradient vector containing the partial derivatives.
Reimplemented from irtkImageFreeFormRegistration.