#include <itkConvolutionImageFilter.h>

Inheritance diagram for itk::ConvolutionImageFilter< TInputImage, TKernelImage, TOutputImage >:

Collaboration diagram for itk::ConvolutionImageFilter< TInputImage, TKernelImage, TOutputImage >:

Public Types
typedef SmartPointer< const Self >	ConstPointer
typedef TInputImage	InputImageType
typedef InputImageType::IndexType	InputIndexType
typedef InputImageType::PixelType	InputPixelType
typedef InputImageType::RegionType	InputRegionType
typedef InputImageType::SizeType	InputSizeType
typedef TKernelImage	KernelImageType
typedef KernelImageType::IndexType	KernelIndexType
typedef KernelImageType::PixelType	KernelPixelType
typedef KernelImageType::RegionType	KernelRegionType
typedef KernelImageType::SizeType	KernelSizeType
typedef TOutputImage	OutputImageType
typedef OutputImageType::IndexType	OutputIndexType
typedef OutputImageType::PixelType	OutputPixelType
typedef OutputImageType::RegionType	OutputRegionType
typedef OutputImageType::SizeType	OutputSizeType
typedef SmartPointer< Self >	Pointer
typedef ConvolutionImageFilter	Self
typedef ConvolutionImageFilterBase < TInputImage, TOutputImage >	Superclass
Public Member Functions
virtual ::itk::LightObject::Pointer	CreateAnother (void) const
virtual const char *	GetNameOfClass () const
Static Public Member Functions
static Pointer	New ()
Static Public Attributes
static const unsigned int	ImageDimension = TInputImage::ImageDimension
Protected Member Functions
	ConvolutionImageFilter ()
void	GenerateData ()
virtual void	GenerateInputRequestedRegion ()
bool	GetKernelNeedsPadding () const
KernelSizeType	GetKernelPadSize () const
template<class TImage >
KernelSizeType	GetKernelRadius (const TImage *kernelImage) const
	~ConvolutionImageFilter ()
Private Member Functions
template<class TImage >
void	ComputeConvolution (const TImage kernelImage, ProgressAccumulator progress)
	ConvolutionImageFilter (const Self &)
void	operator= (const Self &)

Detailed Description

template<class TInputImage, class TKernelImage = TInputImage, class TOutputImage = TInputImage>
class itk::ConvolutionImageFilter< TInputImage, TKernelImage, TOutputImage >

Convolve a given image with an arbitrary image kernel.

This filter operates by centering the flipped kernel at each pixel in the image and computing the inner product between pixel values in the image and pixel values in the kernel. The center of the kernel is defined as $\lfloor (2*i+s-1)/2 \rfloor$ where $i$ is the index and $s$ is the size of the largest possible region of the kernel image. For kernels with odd sizes in all dimensions, this corresponds to the center pixel. If a dimension of the kernel image has an even size, then the center index of the kernel in that dimension will be the largest integral index that is less than the continuous index of the image center.

The kernel can optionally be normalized to sum to 1 using NormalizeOn(). Normalization is off by default.

Warning:: This filter ignores the spacing, origin, and orientation of the kernel image and treats them as identical to those in the input image.

This code was contributed in the Insight Journal paper:

"Image Kernel Convolution" by Tustison N., Gee J. http://hdl.handle.net/1926/1323 http://www.insight-journal.org/browse/publication/208

Author:: Nicholas J. Tustison; James C. Gee

Wiki Examples:

All Examples

Convolve an image with a kernel

Definition at line 65 of file itkConvolutionImageFilter.h.