SoConvolutionWithImageProcessing2d Class Reference
[Frequency Domain]

SoConvolutionWithImageProcessing2d engine More...

#include <ImageViz/Engines/ImageFiltering/FrequencyDomain/SoConvolutionWithImageProcessing2d.h>

Inheritance diagram for SoConvolutionWithImageProcessing2d:

Public Member Functions
	SoConvolutionWithImageProcessing2d ()
Public Attributes
SoSFImageDataAdapter	inImage
SoSFImageDataAdapter	inKernelImage
SoSFBool	autoScale
SoImageVizEngineOutput < SoSFImageDataAdapter, SoImageDataAdapter * >	outImage

Detailed Description

SoConvolutionWithImageProcessing2d engine

Wherever the formulae below refer to $K$ , it is the normalization or automatic scaling coefficient. Set $K$ to 1 if you do not wish to normalize the result.

$\mbox{and}~ K=\sum_{i=-p}^{p}\sum_{j=-q}^{q}|k(i,j)|~\mbox{otherwise}$

The SoConvolutionWithImageProcessing2d filter produces a convolution of image $I$ with another image $k$ . This image $k$ will be converted into a 1D-array and convolution by this general kernel will be performed. It is only implemented for square images.

If image $k$ is of size $(2p+1)\times(2q+1)$ , it is in fact a 2D-array with $(2p+1)$ columns and $(2q+1)$ lines. It will be converted in a 1D-array of $(2p+1)\times(2q+1)$ elements by placing side by side the lines of the $k$ array.