com.openinventor.imageviz.engines.imagefiltering.smoothinganddenoising

Class SoNonLocalMeansFilterProcessing

java.lang.Object

com.openinventor.inventor.Inventor

com.openinventor.inventor.misc.SoBase

com.openinventor.inventor.fields.SoFieldContainer

com.openinventor.inventor.engines.SoEngine

com.openinventor.imageviz.engines.SoImageVizEngine

com.openinventor.imageviz.engines.imagefiltering.smoothinganddenoising.SoNonLocalMeansFilterProcessing

All Implemented Interfaces:

SafeDisposable

public class SoNonLocalMeansFilterProcessing
extends SoImageVizEngine

SoNonLocalMeansFilterProcessing engine The SoNonLocalMeansFilterProcessing engine implements the non-local means algorithm for image denoising.

The algorithm computes a weighted mean of a large set of voxels (search windows) around the voxel to be denoised (target voxel). The voxel weight in the search window is a function of the similarity between this voxel and the target voxel. A voxel very similar to the target voxel will be assigned to a weight close to 1.0. A voxel very different from the target voxel will be assigned to a weight close to 0 and will have a small impact on the final denoised value.

The similarity between a voxel and the target voxel is based on the squared intensity differences between 2 patches of size patchSize centered on the current voxel and the target voxel.

Images are often corrupted by noise because of acquisition process. For an image defined on which is the image domain, the noise added by acquisition is assimilated to an additive white gaussian noise: where

The denoised image is an estimation of as following: with where:

• is the similarity criterion (sum of squared differences)
• is the similarity
• is the indicator function of the search region
• and is the indicator function of the patch region.

The algorithm looks for matches within the search window area around each voxel.

The naive implementation of previous equation has a complexity of that can be reduced to where N*M*Q is the dimension of the input image, P is the patchSize and S is the kernelSize.

File format/default:

NonLocalMeansFilterProcessing {

computeMode	MODE_AUTO
inImage	NULL
kernelShape	BALL
kernelSize	6
patchSize	3
similarity	0.6f

}

Library references: non_local_mean non_local_mean3d

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SoNonLocalMeansFilterProcessing.KernelShapes The shape of the kernel.

Nested classes/interfaces inherited from class com.openinventor.imageviz.engines.SoImageVizEngine

SoImageVizEngine.ComputeModes, SoImageVizEngine.EventArg, SoImageVizEngine.Neighborhood3ds

Nested classes/interfaces inherited from class com.openinventor.inventor.Inventor

Inventor.ConstructorCommand

Field Summary

Fields

Modifier and Type	Field and Description
static int	BALL Deprecated. Use SoNonLocalMeansFilterProcessing.KernelShapes.BALL instead.
SoSFEnum<SoImageVizEngine.ComputeModes>	computeMode Select the compute Mode (2D or 3D or AUTO) .
static int	CUBE Deprecated. Use SoNonLocalMeansFilterProcessing.KernelShapes.CUBE instead.
SoSFImageDataAdapter	inImage Input image.
SoSFEnum<SoNonLocalMeansFilterProcessing.KernelShapes>	kernelShape The search window shape to apply.
SoSFInt32	kernelSize The search window size to apply.
SoImageVizEngineOutput<SoSFImageDataAdapter,SoImageDataAdapter>	outImage The output image.
SoSFInt32	patchSize The patch box size to apply.
SoSFFloat	similarity The similarity is used to compute weight assigned to each voxel in the search window (see expression of ).

Fields inherited from class com.openinventor.imageviz.engines.SoImageVizEngine

CONNECTIVITY_18, CONNECTIVITY_26, CONNECTIVITY_6, MODE_2D, MODE_3D, MODE_AUTO, onBegin, onEnd, onProgress

Fields inherited from class com.openinventor.inventor.Inventor

VERBOSE_LEVEL, ZeroHandle

Constructor Summary

Constructors

Constructor and Description
SoNonLocalMeansFilterProcessing() Constructor.

Method Summary

Methods inherited from class com.openinventor.imageviz.engines.SoImageVizEngine

abortEvaluate, isEvaluating, startEvaluate, waitEvaluate

Methods inherited from class com.openinventor.inventor.engines.SoEngine

copy, getByName, getOutput, getOutputName

Methods inherited from class com.openinventor.inventor.fields.SoFieldContainer

copyFieldValues, copyFieldValues, enableNotify, fieldsAreEqual, get, getAllFields, getEventIn, getEventOut, getField, getFieldName, hasDefaultValues, isNotifyEnabled, set, setToDefaults

Methods inherited from class com.openinventor.inventor.misc.SoBase

dispose, getEXTERNPROTO, getName, getPROTO, isDisposable, isSynchronizable, setName, setSynchronizable, touch

Methods inherited from class com.openinventor.inventor.Inventor

getAddress, getNativeResourceHandle, startInternalThreads, stopInternalThreads

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

CUBE

@Deprecated
public static final int CUBE

Deprecated. Use SoNonLocalMeansFilterProcessing.KernelShapes.CUBE instead.

BALL

@Deprecated
public static final int BALL

Deprecated. Use SoNonLocalMeansFilterProcessing.KernelShapes.BALL instead.

computeMode

public final SoSFEnum<SoImageVizEngine.ComputeModes> computeMode

Select the compute Mode (2D or 3D or AUTO) . Default is MODE_AUTO

inImage

public final SoSFImageDataAdapter inImage

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

kernelShape

public final SoSFEnum<SoNonLocalMeansFilterProcessing.KernelShapes> kernelShape

The search window shape to apply. The shape of the window search area can be use to improve the run time. Indeed, choosing a ball shape reduces of 50% the computation time (in 3D). . Default is BALL

kernelSize

public final SoSFInt32 kernelSize

The search window size to apply.

• In case of a square/cube, a value N will produce a square/cube with 2N+1 side length
• In case of a disk/ball, a value N will produce a disk/ball with a 2N+1 diameter's length The larger the search window is, the better the results usually are. But the size of the search window also effects the run time significantly. This value has to be set to a large enough value so that similar structures can be found within the search window area. Too small values will result in simple blurring of the image because there is not enough structural data within the search window area. Default value is 6.

patchSize

public final SoSFInt32 patchSize

The patch box size to apply. The weight of a voxel in the search windows is computed by comparing the neighborhood of this voxel with the neighborhood of the target voxel. The value represents the edge size of the neighborhood volume in number of voxels (neighborhood is a cube), and affects the quality of the result. If this value is either much smaller or much larger than fine structures in the data the algorithm shows little or no effect at all. This parameter has almost no effect on the computation time. Default value is 3.

similarity

public final SoSFFloat similarity

The similarity is used to compute weight assigned to each voxel in the search window (see expression of ). The squared similarity is proportional to the standard deviation of the assumed gaussian noise of the input image. As result, the larger the value, the more the resulting image will be smoothed. The similarity has no effect on the computation time. Default value is 0.6f.

outImage

public final SoImageVizEngineOutput<SoSFImageDataAdapter,SoImageDataAdapter> outImage

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

Constructor Detail

SoNonLocalMeansFilterProcessing

public SoNonLocalMeansFilterProcessing()

Constructor.