[ VIGRA Homepage | Function Index | Class Index | Namespaces | File List | Main Page ]
FFTWCorrelatePlan< N, Real > Class Template Reference |
#include <vigra/multi_fft.hxx>
Public Member Functions | |
template<class C1 , class C2 , class C3 > | |
void | execute (MultiArrayView< N, Real, C1 > in, MultiArrayView< N, Real, C2 > kernel, MultiArrayView< N, Real, C3 > out) |
Execute a plan to correlate a real array with a real kernel. More... | |
FFTWCorrelatePlan () | |
Create an empty plan. More... | |
template<class C1 , class C2 , class C3 > | |
FFTWCorrelatePlan (MultiArrayView< N, Real, C1 > in, MultiArrayView< N, Real, C2 > kernel, MultiArrayView< N, Real, C3 > out, unsigned int planner_flags=FFTW_ESTIMATE) | |
Create a plan to correlate a real array with a real kernel. More... | |
template<class C1 , class C2 , class C3 > | |
FFTWCorrelatePlan (Shape inOut, Shape kernel, bool useFourierKernel=false, unsigned int planner_flags=FFTW_ESTIMATE) | |
Create a plan from just the shape information. More... | |
template<class C1 , class C2 , class C3 > | |
void | init (MultiArrayView< N, Real, C1 > in, MultiArrayView< N, Real, C2 > kernel, MultiArrayView< N, Real, C3 > out, unsigned int planner_flags=FFTW_ESTIMATE) |
Init a plan to convolve a real array with a real kernel. More... | |
Like FFTWConvolvePlan, but performs correlation rather than convolution.
See vigra::FFTWConvolvePlan for details.
Usage:
#include <vigra/multi_fft.hxx>
Namespace: vigra
Create an empty plan.
The plan can be initialized later by one of the init() functions.
FFTWCorrelatePlan | ( | MultiArrayView< N, Real, C1 > | in, |
MultiArrayView< N, Real, C2 > | kernel, | ||
MultiArrayView< N, Real, C3 > | out, | ||
unsigned int | planner_flags = FFTW_ESTIMATE |
||
) |
Create a plan to correlate a real array with a real kernel.
The kernel must be defined in the spatial domain. See correlateFFT() for detailed information on required shapes and internal padding.
FFTW_ESTIMATE
will guess optimal algorithm settings or read them from pre-loaded "wisdom". FFTWCorrelatePlan | ( | Shape | inOut, |
Shape | kernel, | ||
bool | useFourierKernel = false , |
||
unsigned int | planner_flags = FFTW_ESTIMATE |
||
) |
Create a plan from just the shape information.
See convolveFFT() for detailed information on required shapes and internal padding.
FFTW_ESTIMATE
will guess optimal algorithm settings or read them from pre-loaded "wisdom". void init | ( | MultiArrayView< N, Real, C1 > | in, |
MultiArrayView< N, Real, C2 > | kernel, | ||
MultiArrayView< N, Real, C3 > | out, | ||
unsigned int | planner_flags = FFTW_ESTIMATE |
||
) |
Init a plan to convolve a real array with a real kernel.
See the constructor with the same signature for details.
void execute | ( | MultiArrayView< N, Real, C1 > | in, |
MultiArrayView< N, Real, C2 > | kernel, | ||
MultiArrayView< N, Real, C3 > | out | ||
) |
Execute a plan to correlate a real array with a real kernel.
The array shapes must be the same as in the corresponding init function or constructor. However, execute() can be called several times on the same plan, even with different arrays, as long as they have the appropriate shapes.
© Ullrich Köthe (ullrich.koethe@iwr.uni-heidelberg.de) |
html generated using doxygen and Python
|