SoMedianFilterProcessing Class Reference
[Smoothing And Denoising]

#include <ImageViz/Engines/ImageFiltering/SmoothingAndDenoising/SoMedianFilterProcessing.h>

Inheritance diagram for SoMedianFilterProcessing:

Public Types
enum	KernelMode { ITERATIVE = 0, CUBE = 1, BALL = 2 }
Public Member Functions
	SoMedianFilterProcessing ()
Public Attributes
SoSFEnum	computeMode
SoSFEnum	neighborhood3d
SoSFImageDataAdapter	inImage
SoSFInt32	elementSize
SoSFEnum	kernelMode
SoImageVizEngineOutput < SoSFImageDataAdapter, SoImageDataAdapter * >	outImage

Detailed Description

SoMedianFilterProcessing engine

The SoMedianFilterProcessing engine smoothes the image using morphological operators.

For an introduction to image filters, see section Images Filters.

This module uses morphological operators to set the pixel value to the median for the defined neighborhood. The SoMedianFilterProcessing usually works well when images contain non-Gaussian noise and/or very small artifacts. It does not cause blurring to the same extend as the SoBoxFilterProcessing, but does take considerably longer to execute.

The gray levels of all pixels in the neighborhood are sorted from the smallest value to the largest one. The central pixel in the sort is then the median value, i.e. the value for which there are as many lower gray levels as higher ones. The process may be iterated.

To explain this, consider the array:

$\left[\begin{array}{ccc} 12 & 17 & 15 \\ 20 & 14 & 16 \\ 18 & 19 & 14 \end{array}\right]$

The sorted gray level values will be: 12 14 14 15 17 18 19 20.
The median value is 16, and it will be the value of the output pixel.

The Median filtering considers the lines and then the columns.

FILE FORMAT/DEFAULT

computeMode	MODE_AUTO
neighborhood3d	CONNECTIVITY_26
inImage	NULL
elementSize	3
kernelMode	CUBE

Library references: medianfilter medianfilter3d

See related examples:: InteractiveFiltering, InteractiveFloodFill, DenoisingImage2D

Member Enumeration Documentation

enum SoMedianFilterProcessing::KernelMode

Enumerator:

ITERATIVE

This mode performs an iterative median filtering of the input image.

In 2D: iteratively, a pixel is assigned to the median value in a 3x3 neighborhood around the corresponding pixel in the previous iteration. The same goes in 3D but the neighborhood is controlled by the 3D connectivity (6-, 18- or 26-direct neighbors). In ITERATIVE mode the elementSize field represents the number of iterations. The ITERATIVE mode is recommended to be used with 2D image with an elementSize lower than 3. For greater values, ITERATIVE mode is faster than SQUARE mode but doesn't change significantly the results, while SQUARE mode applies the filter in the effective window size.

CUBE

This mode specifies a square neighborhood in 2D or a cube neighborhood in 3D.

In 2D, it will produce a square of side 2 x elementSize + 1 pixels. In 3D, it will produce a cube with of size 2 x elementSize + 1 voxels.

BALL

This mode specifies a disk neighborhood in 2D or a sphere neighborhood in 3D.

In 2D, it will produce a disk with of diameter 2 x + 1 pixels. In 3D, it will produce a spshee of diameter 2 x elementSize + 1 voxels.

Constructor & Destructor Documentation

SoMedianFilterProcessing::SoMedianFilterProcessing ( )

Constructor.

Member Data Documentation

SoSFEnum SoMedianFilterProcessing::computeMode

Select the compute Mode (2D or 3D or AUTO) Use enum ComputeMode.

Default is MODE_AUTO

SoSFInt32 SoMedianFilterProcessing::elementSize

The half size of the structuring element.

A structuring element always has an odd side length (3x3, 5x5, etc) which is defined by 2 * elementSize + 1. Default value is 3.

SoSFImageDataAdapter SoMedianFilterProcessing::inImage

The input grayscale image.

Default value is NULL. Supported types include: grayscale binary label image.

SoSFEnum SoMedianFilterProcessing::kernelMode

The shape of the 3D neighborhood.

Use enum KernelMode. Default is CUBE

SoSFEnum SoMedianFilterProcessing::neighborhood3d

In 3D configuration (see computeMode), the neighborhood connectivity defines the connectivity considered for processing adjacent voxels.

Use enum Neighborhood3d. Default is CONNECTIVITY_26.

SoImageVizEngineOutput<SoSFImageDataAdapter,SoImageDataAdapter*> SoMedianFilterProcessing::outImage

The output image.

Default value is NULL. Supported types include: grayscale binary label color image.

The documentation for this class was generated from the following file:

ImageViz/Engines/ImageFiltering/SmoothingAndDenoising/SoMedianFilterProcessing.h

SoMedianFilterProcessing Class Reference [Smoothing And Denoising]

Public Types

Public Member Functions