#include <itkWienerDeconvolutionImageFilter.h>

Inheritance diagram for itk::WienerDeconvolutionImageFilter< TInputImage, TKernelImage, TOutputImage, TInternalPrecision >:

Collaboration diagram for itk::WienerDeconvolutionImageFilter< TInputImage, TKernelImage, TOutputImage, TInternalPrecision >:

List of all members.

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef TInputImage	InputImageType
typedef Superclass::InputIndexType	InputIndexType
typedef Superclass::InputPixelType	InputPixelType
typedef Superclass::InputRegionType	InputRegionType
typedef Superclass::InputSizeType	InputSizeType
typedef Superclass::InternalComplexImagePointerType	InternalComplexImagePointerType
typedef Superclass::InternalComplexImageType	InternalComplexImageType
typedef Superclass::InternalComplexType	InternalComplexType
typedef Superclass::InternalImagePointerType	InternalImagePointerType
typedef Superclass::InternalImageType	InternalImageType
typedef TKernelImage	KernelImageType
typedef Superclass::KernelIndexType	KernelIndexType
typedef Superclass::KernelPixelType	KernelPixelType
typedef Superclass::KernelRegionType	KernelRegionType
typedef Superclass::KernelSizeType	KernelSizeType
typedef TOutputImage	OutputImageType
typedef Superclass::OutputIndexType	OutputIndexType
typedef Superclass::OutputPixelType	OutputPixelType
typedef Superclass::OutputRegionType	OutputRegionType
typedef Superclass::OutputSizeType	OutputSizeType
typedef SmartPointer< Self >	Pointer
typedef WienerDeconvolutionImageFilter	Self
typedef Superclass::SizeValueType	SizeValueType
typedef FFTConvolutionImageFilter < TInputImage, TOutputImage >	Superclass
Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother (void) const
virtual const char *	GetNameOfClass () const

virtual void	SetNoiseVariance (double _arg)
virtual double	GetNoiseVariance () const
Static Public Member Functions
static Pointer	New ()
Static Public Attributes
static const unsigned int	ImageDimension = TInputImage::ImageDimension
Protected Member Functions
void	GenerateData ()
virtual void	PrintSelf (std::ostream &os, Indent indent) const
	WienerDeconvolutionImageFilter ()
	~WienerDeconvolutionImageFilter ()
Private Member Functions
void	operator= (const Self &)
	WienerDeconvolutionImageFilter (const Self &)
Private Attributes
double	m_NoiseVariance

Detailed Description

template<class TInputImage, class TKernelImage = TInputImage, class TOutputImage = TInputImage, class TInternalPrecision = double>
class itk::WienerDeconvolutionImageFilter< TInputImage, TKernelImage, TOutputImage, TInternalPrecision >

The Wiener deconvolution image filter is designed to restore an image convolved with a blurring kernel while keeping noise enhancement to a minimum.

The Wiener filter aims to minimize noise enhancement induced by frequencies with low signal-to-noise ratio. The Wiener filter kernel is defined in the frequency domain as $W(\omega) = H^*(\omega) / (|H(\omega)|^2 + (1 / SNR(\omega)))$ where $H(\omega)$ is the Fourier transform of the blurring kernel with which the original image was convolved and the signal-to-noise ratio $SNR(\omega)$ . $SNR(\omega)$ is defined by $P_f(\omega) / P_n(\omega)$ where $P_f(\omega)$ is the power spectral density of the uncorrupted signal and $P_n(\omega)$ is the power spectral density of the noise. When applied to the input blurred image, this filter produces an estimate $\hat{f}(x)$ of the true underlying signal $f(x)$ that minimizes the expected error between $\hat{f}(x)$ and $f(x)$ .

This filter requires two inputs, the image to be deconvolved and the blurring kernel. These two inputs can be set using the methods SetInput() and SetKernelImage(), respectively.

The power spectral densities of the signal and noise are typically unavailable for a given problem. In particular, $P_f(\omega)$ cannot be computed from $f(x)$ because this unknown signal is precisely the signal that this filter aims to recover. Nevertheless, it is common for the noise to have a power spectral density that is flat or decreasing significantly more slowly than the power spectral density of a typical image as the frequency $\omega$ increases. Hence, $P_n(\omega)$ can typically be approximated with a constant, and this filter makes this assumption (see the NoiseVariance member variable). $P_f(\omega)$ , on the other hand, will vary with input. This filter computes the power spectral density of the input blurred image, subtracts the power spectral density of the noise, and uses the result as the estimate of $P_f(\omega)$ .

For further information on the Wiener deconvolution filter, please see "Digital Signal Processing" by Kenneth R. Castleman, Prentice Hall, 1995

Author:: Gaeten Lehmann, Biologie du Developpement et de la Reproduction, INRA de Jouy-en-Josas, France; Chris Mullins, The University of North Carolina at Chapel Hill; Cory Quammen, The University of North Carolina at Chapel Hill

Definition at line 75 of file itkWienerDeconvolutionImageFilter.h.